guide/gpl-lgpl.tex

4955 lines
269 KiB
TeX
Raw Blame History

This file contains invisible Unicode characters

This file contains invisible Unicode characters that are indistinguishable to humans but may be processed differently by a computer. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

% gpl-lgpl.tex -*- LaTeX -*-
% Tutorial Text for the Detailed Study and Analysis of GPL and LGPL course
%
% Copyright (C) 2003, 2004, 2005, 2006 Free Software Foundation, Inc.
% Copyright (C) 2014 Bradley M. Kuhn
% License: CC-By-SA-4.0
% The copyright holders hereby grant the freedom to copy, modify, convey,
% Adapt, and/or redistribute this work under the terms of the Creative
% Commons Attribution Share Alike 4.0 International License.
% This text is distributed in the hope that it will be useful, but
% WITHOUT ANY WARRANTY; without even the implied warranty of
% MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
% You should have received a copy of the license with this document in
% a file called 'CC-By-SA-4.0.txt'. If not, please visit
% https://creativecommons.org/licenses/by-sa/4.0/legalcode to receive
% the license text.
% FIXME-LATER: I should make a macro like the Texinfo @xref stuff for places
% where I'm saying ``see section X in this tutorial'', so that the extra
% verbiage isn't there in the HTML versions that I'll eventually do.
% Maybe something like that already exists? In the worst case, I could
% adapt @xref from texinfo.texi for it.
\newcommand{\defn}[1]{\emph{#1}}
\part{Detailed Analysis of the GNU GPL and Related Licenses}
{\parindent 0in
\tutorialpartsplit{``Detailed Analysis of the GNU GPL and Related Licenses''}{This part} is: \\
\begin{tabbing}
Copyright \= \copyright{} 2003--2007 \hspace{.1mm} \= \kill
Copyright \> \copyright{} 2014 \> Bradley M. Kuhn \\
Copyright \> \copyright{} 2014 \> Anthony K. Sebro, Jr. \\
Copyright \> \copyright{} 2003--2007 \> Free Software Foundation, Inc.
\end{tabbing}
\vspace{1in}
\begin{center}
Authors of \tutorialpartsplit{``Detailed Analysis of the GNU GPL and Related Licenses''}{this part} are: \\
Free Software Foundation, Inc. \\
Bradley M. Kuhn \\
David ``Novalis'' Turner \\
Daniel B. Ravicher \\
Tony Sebro \\
John Sullivan
\vspace{.3in}
The copyright holders of \tutorialpartsplit{``Detailed Analysis of the GNU GPL and Related Licenses''}{this part} hereby grant the freedom to copy, modify,
convey, Adapt, and/or redistribute this work under the terms of the Creative
Commons Attribution Share Alike 4.0 International License. A copy of that
license is available at
\verb=https://creativecommons.org/licenses/by-sa/4.0/legalcode=.
\end{center}
}
\bigskip
\bigskip
\tutorialpartsplit{This tutorial}{This part of the tutorial} gives a
comprehensive explanation of the most popular Free Software copyright
license, the GNU General Public License (``GNU GPL'', or sometimes just
``GPL'') -- both version 2 (``GPLv2'') and version 3 (``GPLv3'') -- and
teaches lawyers, software developers, managers and business people how to use
the GPL (and GPL'd software) successfully both as a community-building
``Constitution'' for a software project, and to incorporate copylefted
software into a new Free Software business and in existing, successful
enterprises.
To successfully benefit of from this part of the tutorial, readers should
have a general familiarity with software development processes. A basic
understanding of how copyright law applies to software is also helpful. The
tutorial is of most interest to lawyers, software developers and managers who
run or advise software businesses that modify and/or redistribute software
under the terms of the GNU GPL (or who wish to do so in the future), and those
who wish to make use of existing GPL'd software in their enterprise.
Upon completion of this part of the tutorial, successful readers can expect
to have learned the following:
\begin{itemize}
\item The freedom-defending purpose of various terms in the GNU GPLv2 and GPLv3.
\item The differences between GPLv2 and GPLv3.
\item The redistribution options under the GPLv2 and GPLv3.
\item The obligations when modifying GPLv2'd or GPLv3'd software.
\item How to build a plan for proper and successful compliance with the GPL.
\item The business advantages that the GPL provides.
\item The most common business models used in conjunction with the GPL.
\item How existing GPL'd software can be used in existing enterprises.
\item The basics of LGPLv2.1 and LGPLv3, and how they
differs from the GPLv2 and GPLv3, respectively.
\item The basics to begin understanding the complexities regarding
derivative and combined works of software.
\end{itemize}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% END OF ABSTRACTS SECTION
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% START OF DAY ONE COURSE
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\chapter{What Is Software Freedom?}
Study of the GNU General Public License (herein, abbreviated as \defn{GNU
GPL} or just \defn{GPL}) must begin by first considering the broader world
of software freedom. The GPL was not created in a vacuum. Rather, it was
created to embody and defend a set of principles that were set forth at the
founding of the GNU project and the Free Software Foundation (FSF) -- the
preeminent organization that upholds, defends and promotes the philosophy of software
freedom. A prerequisite for understanding both of the popular versions
of the GPL
(GPLv2 and GPLv3) and their terms and conditions is a basic understanding of
the principles behind them. The GPL family of licenses are unlike nearly all
other software licenses in that they are designed to defend and uphold these
principles.
\section{The Free Software Definition}
\label{Free Software Definition}
The Free Software Definition is set forth in full on FSF's website at
\verb0http://fsf.org/0 \verb0philosophy/free-sw.html0. This section presents
an abbreviated version that will focus on the parts that are most pertinent
to the GPL\@.
A particular program grants software freedom to a particular user if that
user is granted the following freedoms:
\begin{itemize}
\item The freedom to run the program, for any purpose.
\item The freedom to study how the program works, and modify it
\item The freedom to redistribute copies.
\item The freedom to distribute copies of modified versions to others.
\end{itemize}
The focus on ``a particular user'' is particularly pertinent here. It is not
uncommon for the same version of a specific program to grant these freedoms
to some subset of its user base, while others have none or only some of these
freedoms. Section~\ref{Proprietary Relicensing} talks in detail about how
this can unfortunately happen even if a program is released under the GPL\@.
Many people refer to software that gives these freedoms as ``Open Source.''
Besides having a different political focus than those who call it Free
Software,\footnote{The political differences between the Free Software
Movement and the Open Source Movement are documented on FSF's Web site at
{\tt http://www.fsf.org/licensing/essays/free-software-for-freedom.html}.}
Those who call the software ``Open Source'' are often focused on a side
issue. Specifically, user access to the source code of a program is a
prerequisite to make use of the freedom to modify. However, the important
issue is what freedoms are granted in the license of that source code.
Software freedom is only complete when no restrictions are imposed on how
these freedoms are exercised. Specifically, users and programmers can
exercise these freedoms noncommercially or commercially. Licenses that grant
these freedoms for noncommercial activities but prohibit them for commercial
activities are considered non-free. Even the Open Source Initiative
(\defn{OSI}) (the arbiter of what is considered ``Open Source'') also rules
such licenses not in fitting with its ``Open Source Definition''.
In general, software for which most or all of these freedoms are
restricted in any way is called ``non-Free Software.'' Typically, the
term ``proprietary software'' is used more or less interchangeably with
``non-Free Software.'' Personally, I tend to use the term ``non-Free
Software'' to refer to noncommercial software that restricts freedom
(such as ``shareware'') and ``proprietary software'' to refer to
commercial software that restricts freedom (such as nearly all of
Microsoft's and Oracle's offerings).
Keep in mind that the none of the terms ``software freedom'', ``open source''
and ``free software'' are known to be trademarked or otherwise legally
restricted by any organization in
any jurisdiction. As such, it's quite common that these terms are abused and
misused by parties who wish to bank on the popularity of software freedom.
When one considers using, modifying or redistributing a software package that
purports to be Open Source or Free Software, one \textbf{must} verify that
the license grants software freedom.
Furthermore, throughout this text, we generally prefer the term ``software
freedom'', as this is the least ambiguous term available to describe software
that meets the Free Software Definition. For example, it is well known and
often discussed that the adjective ``free'' has two unrelated meanings in
English: ``free as in freedom'' and ``free as in price''. Meanwhile, the
term ``open source'' is even more confusing, because it appears to refer only to the
``freedom to study'', which is merely a subset of one of the four freedoms.
The remainder of this section considers each of each component of software
freedom in detail.
\subsection{The Freedom to Run}
\label{freedom-to-run}
The first tenet of software freedom is the user's fully unfettered right to
run the program. The software's license must permit any conceivable use of
the software. Perhaps, for example, the user has discovered an innovative
use for a particular program, one that the programmer never could have
predicted. Such a use must not be restricted.
It was once rare that this freedom was restricted by even proprietary
software; but such is quite common today. Most End User License Agreements
(EULAs) that cover most proprietary software typically restrict some types of
uses. Such restrictions of any kind are an unacceptable restriction on
software freedom.
\subsection{The Freedom to Change and Modify}
Perhaps the most useful right of software freedom is the users' right to
change, modify and adapt the software to suit their needs. Access to the
source code and related build and installation scripts are an essential part
of this freedom. Without the source code, and the ability to build and
install the binary applications from that source, users cannot effectively
exercise this freedom.
Programmers directly benefit from this freedom. However, this freedom
remains important to users who are not programmers. While it may seem
counterintuitive at first, non-programmer users often exercise this freedom
indirectly in both commercial and noncommercial settings. For example, users
often seek noncommercial help with the software on email lists and in user
groups. To make use of such help they must either have the freedom to
recruit programmers who might altruistically assist them to modify their
software, or to at least follow rote instructions to make basic modifications
themselves.
More commonly, users also exercise this freedom commercially. Each user, or
group of users, may hire anyone they wish in a competitive free market to
modify and change the software. This means that companies have a right to
hire anyone they wish to modify their Free Software. Additionally, such
companies may contract with other companies to commission software
modification.
\subsection{The Freedom to Copy and Share}
Users share Free Software in a variety of ways. Software freedom advocates
work to eliminate a fundamental ethical dilemma of the software age: choosing
between obeying a software license and friendship (by giving away a copy of a
program to your friend who likes the software you are using). Licenses that
respect software freedom, therefore, permit altruistic sharing of software
among friends.
The commercial environment also benefits of this freedom. Commercial sharing
includes selling copies of Free Software: that is, Free Software can
be distribted for any monetary
price to anyone. Those who redistribute Free Software commercially also have
the freedom to selectively distribute (i.e., you can pick your customers) and
to set prices at any level that redistributor sees fit.
Of course, most people get copies of Free Software very cheaply (and
sometimes without charge). The competitive free market of Free Software
tends to keep prices low and reasonable. However, if someone is willing to
pay billions of dollars for one copy of the GNU Compiler Collection, such a
sale is completely permitted.
Another common instance of commercial sharing is service-oriented
distribution. For example, some distribution vendors provide immediate
security and upgrade distribution via a special network service. Such
distribution is not necessarily contradictory with software freedom.
(Section~\ref{Business Models} of this tutorial talks in detail about some
common Free Software business models that take advantage of the freedom to
share commercially.)
\subsection{The Freedom to Share Improvements}
The freedom to modify and improve is somewhat empty without the freedom to
share those improvements. The Software freedom community is built on the
pillar of altruistic sharing of improved Free Software. Historically
it was typical for a
Free Software project to sprout a mailing list where improvements
would be shared
freely among members of the development community\footnote{This is still
commonly the case, though today there are other or additional ways of
sharing Free Software.}. Such noncommercial
sharing is the primary reason that Free Software thrives.
Commercial sharing of modified Free Software is equally important.
For commercial support to exist in a competitive free market, all
developers -- from single-person contractors to large software
companies -- must have the freedom to market their services as
improvers of Free Software. All forms of such service marketing must
be equally available to all.
For example, selling support services for Free Software is fully
permitted. Companies and individuals can offer themselves as ``the place
to call'' when software fails or does not function properly. For such a
service to be meaningful, the entity offering that service needs the
right to modify and improve the software for the customer to correct any
problems that are beyond mere user error.
Software freedom licenses also permit any entity to distribute modified
versions of Free Software. Most Free Software programs have a ``standard
version'' that is made available from the primary developers of the software.
However, all who have the software have the ``freedom to fork'' -- that is,
make available nontrivial modified versions of the software on a permanent or
semi-permanent basis. Such freedom is central to vibrant developer and user
interaction.
Companies and individuals have the right to make true value-added versions
of Free Software. They may use freedom to share improvements to
distribute distinct versions of Free Software with different functionality
and features. Furthermore, this freedom can be exercised to serve a
disenfranchised subset of the user community. If the developers of the
standard version refuse to serve the needs of some of the software's
users, other entities have the right to create a long- or short-lived fork
to serve that sub-community.
\section{How Does Software Become Free?}
The previous section set forth key freedoms and rights that are referred to
as ``software freedom''. This section discusses the licensing mechanisms
used to enable software freedom. These licensing mechanism were ultimately
created as a community-oriented ``answer'' to the existing proprietary
software licensing mechanisms. Thus, first, consider carefully why
proprietary software exists in the first place.
Proprietary software exists at all only because it is governed by copyright
law.\footnote{This statement is admittedly an oversimplification. Patents and
trade secrets can cover software and make it effectively non-Free, and one
can contract away their rights and freedoms regarding software, or source
code can be practically obscured in binary-only distribution without
reliance on any legal system. However, the primary control mechanism for
software is copyright, and therefore this section focuses on how copyright
restrictions make software proprietary.} Copyright law, with respect to
software, typically governs copying, modifying, and redistributing that
software (For details of this in the USA, see
\href{http://www.copyright.gov/title17/92chap1.html#106}{\S~106} and
\href{http://www.copyright.gov/title17/92chap1.html#117}{\S~117} of
\href{http://www.law.cornell.edu/uscode/text/17}{Title 17} of the
\textit{United States Code}).\footnote{Copyright law in general also governs
``public performance'' of copyrighted works. There is no generally agreed
definition for public performance of software and both GPLv2 and GPLv3 do
not restrict public performance.} By law (in the USA and in most other
jurisdictions), the copyright holder (most typically, the author) of the work controls
how others may copy, modify and/or distribute the work. For proprietary
software, these controls are used to prohibit these activities. In addition,
proprietary software distributors further impede modification in a practical
sense by distributing only binary code and keeping the source code of the
software secret.
Copyright is not a natural state, it is a legal construction. In the USA, the
Constitution permits, but does not require, the creation of copyright law as
federal legislation. Software, since it is ``an original works of authorship
fixed in any tangible medium of expression ... from which they can be
perceived, reproduced, or otherwise communicated, either directly or with the
aid of a machine or device'' (as stated in
\href{http://www.law.cornell.edu/uscode/text/17/102}{17 USC \S~102}), is thus
covered by the statute, and is copyrighted by default.
However, software, in its natural state without copyright, is Free
Software. In an imaginary world with no copyright, the rules would be
different. In this world, when you received a copy of a program's source
code, there would be no default legal system to restrict you from sharing it
with others, making modifications, or redistributing those modified
versions.\footnote{Note that this is again an oversimplification; the
complexities with this argument are discussed in
Section~\ref{software-and-non-copyright}.}
Software in the real world is copyrighted by default and is automatically
covered by that legal system. However, it is possible to move software out
of the domain of the copyright system. A copyright holder can often
\defn{disclaim} their copyright. (For example, under USA copyright law
it is possible for a copyright holder to engage in conduct resulting
in abandonment of copyright.) If copyright is disclaimed, the software is
effectively no longer restricted by copyright law. Software not restricted by copyright is in the
``public domain.''
\subsection{Public Domain Software}
In the USA and other countries that
are parties to the Berne Convention on Copyright, software is copyrighted
automatically by the author when she fixes the software in a tangible
medium. In the software world, this usually means typing the source code
of the software into a file.
Imagine if authors could truly disclaim those default control of copyright
law. If so, the software is in the public domain --- no longer covered by
copyright. Since copyright law is the construction allowing for most
restrictions on software (i.e., prohibition of copying, modification, and
redistribution), removing the software from the copyright system usually
yields software freedom for its users.
Carefully note that software truly in the public domain is \emph{not} licensed
in any way. It is confusing to say software is ``licensed for the
public domain,'' or any phrase that implies the copyright holder gave
express permission to take actions governed by copyright law.
Copyright holders who state that they are releasing their code into
the public domain are effectively renouncing copyright controls on
the work. The law gave the copyright holders exclusive controls over the
work, and they chose to waive those controls. Software that is, in
this sense, in the public domain
is conceptualized by the developer as having no copyright and thus no license. The software freedoms discussed in
Section~\ref{Free Software Definition} are all granted because there is no
legal system in play to take them away.
Admittedly, a discussion of public domain software is an oversimplified
example.
Because copyright controls are usually automatically granted and because, in
some jurisdictions, some copyright controls cannot be waived (see
Section~\ref{non-usa-copyright} for further discussion), many copyright
holders sometimes incorrectly believe a work has been placed in the public
domain. Second, due to aggressive lobbying by the entertainment industry,
the ``exclusive Right'' of copyright, that was supposed to only exist for
``Limited Times'' according to the USA Constitution, appears to be infinite:
simply purchased on the installment plan rather than in whole. Thus, we must
assume no works of software will fall into the public domain merely due to
the passage of time.
Nevertheless, under USA law it is likely that the typical
disclaimers of copyright or public domain dedications we see in the
Free Software world would be interpreted by courts as copyright
abandonment, leading to a situation in which the user effectively receives a
maximum grant of copyright freedoms, similar to a maximally-permissive
Free Software license.
The best example of software known to truly be in the public domain is software
that is published by the USA government. Under
\href{http://www.law.cornell.edu/uscode/text/17/105}{17 USC 101 \S~105}, all
works published by the USA Government are not copyrightable in the USA.
\subsection{Why Copyright Free Software?}
If simply disclaiming copyright on software yields Free Software, then it
stands to reason that putting software into the public domain is the
easiest and most straightforward way to produce Free Software. Indeed,
some major Free Software projects have chosen this method for making their
software Free. However, most of the Free Software in existence \emph{is}
copyrighted. In most cases (particularly in those of FSF and the GNU
Project), this was done due to very careful planning.
Software released into the public domain does grant freedom to those users
who receive the standard versions on which the original author disclaimed
copyright. However, since the work is not copyrighted, any nontrivial
modification made to the work is fully copyrightable.
Free Software released into the public domain initially is Free, and
perhaps some who modify the software choose to place their work into the
public domain as well. However, over time, some entities will choose to
proprietarize their modified versions. The public domain body of software
feeds the proprietary software. The public commons disappears, because
fewer and fewer entities have an incentive to contribute back to the
commons. They know that any of their competitors can proprietarize their
enhancements. Over time, almost no interesting work is left in the public
domain, because nearly all new work is done by proprietarization.
A legal mechanism is needed to redress this problem. FSF was in fact
originally created primarily as a legal entity to defend software freedom,
and that work of defending software freedom is a substantial part of
its work today. Specifically because of this ``embrace, proprietarize and
extend'' cycle, FSF made a conscious choice to copyright its Free Software,
and then license it under ``copyleft'' terms. Many, including the
developers of the kernel named Linux, have chosen to follow this paradigm.
\label{copyleft-definition}
Copyleft is a legal strategy and mechanism to defend, uphold and propagate software
freedom. The basic technique of copyleft is as follows: copyright the
software, license it under terms that give all the software freedoms, but
use the copyright law controls to ensure that all who receive a copy of
the software have equal rights and freedom. In essence, copyleft grants
freedom, but forbids others to forbid that freedom to anyone else along
the distribution and modification chains.
Copyleft is a general concept. Much like ideas for what a computer might
do must be \emph{implemented} by a program that actually does the job, so
too must copyleft be implemented in some concrete legal structure.
``Share and share alike'' is a phrase that is used often enough to explain the
concept behind copyleft, but to actually make it work in the real world, a
true implementation in legal text must exist. The GPL is the primary
implementation of copyleft in copyright licensing language.
\subsection{Software and Non-Copyright Legal Regimes}
\label{software-and-non-copyright}
The use, modification and distribution of software, like many endeavors,
simultaneously interacts with multiple different legal regimes. As was noted
early via footnotes, copyright is merely the \textit{most common way} to
restrict users' rights to copy, share, modify and/or redistribute software.
However, proprietary software licenses typically use every mechanism
available to subjugate users. For example:
\begin{itemize}
\item Unfortunately, despite much effort by many in the software freedom
community to end patents that read on software (i.e., patents on
computational ideas), they still ultimately exist. As such, a software
program might otherwise be seemly unrestricted, but a patent might read on
the software and ruin everything for its users.\footnote{See
\S\S~\ref{gpl-implied-patent-grant},~\ref{GPLv2s7},~\ref{GPLv3s11} for more
discussion on how the patent system interacts with copyleft, and read
Richard M.~Stallman's essay,
\href{http://www.wired.com/opinion/2012/11/richard-stallman-software-patents/}{\textit{Lets
Limit the Effect of Software Patents, Since We Cant Eliminate Them}}
for more information on the problems these patents present to society.}
\item Digital Restrictions Management (usually called \defn{DRM}) is often
used to impose technological restrictions on users' ability to exercise
software freedom that they might otherwise be granted\footnote{See
\S~\ref{GPLv3-drm} for more information on how GPL deals with this issue.}.
The simplest (and perhaps oldest) form of DRM, of course, is separating
software source code (read by humans), from their compiled binaries (read
only by computers). Furthermore,
\href{http://www.law.cornell.edu/uscode/text/17/1201}{17 USC~\S1201} often
prohibits users legally from circumventing some of these DRM systems.
\item Most EULAs also include a contractual agreement that bind users further
by forcing them to agree to a contractual, prohibitive software license
before ever even using the software.
\end{itemize}
Thus, most proprietary software restricts users via multiple interlocking
legal and technological means. Any license that truly respect the software
freedom of all users must not only grant appropriate copyright permissions,
but also \textit{prevent} restrictions from other legal and technological
means like those listed above.
\subsection{Non-USA Copyright Regimes}
\label{non-usa-copyright}
Generally speaking, copyright law operates similarly enough in countries that
have signed the Berne Convention on Copyright, and software freedom licenses
have generally taken advantage of this international standardization of
copyright law. However, copyright law does differ from country to country,
and commonly, software freedom licenses like GPL must be considered under the
copyright law in the jurisdiction where any licensing dispute occurs.
Those who are most familiar with the USA's system of copyright often are
surprised to learn that there are certain copyright controls that cannot be
waived nor disclaimed. Specifically, many copyright regimes outside the USA
recognize a concept of moral rights of authors. Typically, moral rights are
fully compatible with respecting software freedom, as they are usually
centered around controls that software freedom licenses generally respect,
such as the right of an authors to require proper attribution for their work.
\section{A Community of Equality}
The previous section described the principles of software freedom, a brief
introduction to mechanisms that typically block these freedoms, and the
simplest ways that copyright holders might grant those freedoms to their
users for their copyrighted works of software. The previous section also
introduced the idea of \textit{copyleft}: a licensing mechanism to use
copyright to not only grant software freedom to users, but also to uphold
those rights against those who might seek to curtail them.
Copyleft, as defined in \S~\ref{copyleft-definition}, is a general term this
mechanism. The remainder of this text will discuss details of various
real-world implementations of copyleft -- most notably, the GPL\@.
This discussion begins first with some general explanation of what the GPL is
able to do in software development communities. After that brief discussion
in this section, deeper discussion of how GPL accomplishes this in practice
follows in the next chapter.
Simply put, though, the GPL ultimately creates an community of equality for
both business and noncommercial users.
\subsection{The Noncommercial Community}
A GPL'd code base becomes a center of a vibrant development and user
community. Traditionally, volunteers, operating noncommercially out of
keen interest or ``scratch an itch'' motivations, produce initial versions
of a GPL'd system. Because of the efficient distribution channels of the
Internet, any useful GPL'd system is adopted quickly by noncommercial
users.
Fundamentally, the early release and quick distribution of the software
gives birth to a thriving noncommercial community. Users and developers
begin sharing bug reports and bug fixes across a shared intellectual
commons. Users can trust the developers, because they know that if the
developers fail to address their needs or abandon the project, the GPL
ensures that someone else has the right to pick up development.
Developers know that the users cannot redistribute their software without
passing along the rights granted by GPL, so they are assured that every
one of their users is treated equally.
Because of the symmetry and fairness inherent in GPL'd distribution,
nearly every GPL'd package in existence has a vibrant noncommercial user
and developer base.
\subsection{The Commercial Community}
By the same token, nearly all established GPL'd software systems have a
vibrant commercial community. Nearly every GPL'd system that has gained
wide adoption from noncommercial users and developers eventually begins
to fuel a commercial system around that software.
For example, consider the Samba file server system that allows Unix-like
systems (including GNU/Linux) to serve files to Microsoft Windows systems.
Two graduate students originally developed Samba in their spare time and
it was deployed noncommercially in academic environments\footnote{See
\href{http://turtle.ee.ncku.edu.tw/docs/samba/history}{Andrew Tridgell's
``A bit of history and a bit of fun''}}. However, very
soon for-profit companies discovered that the software could work for them
as well, and their system administrators began to use it in place of
Microsoft Windows NT file-servers. This served to lower the cost of
running such servers by orders of magnitude. There was suddenly room in
Windows file-server budgets to hire contractors to improve Samba. Some of
the first people hired to do such work were those same two graduate
students who originally developed the software.
The noncommercial users, however, were not concerned when these two
fellows began collecting paychecks off of their GPL'd work. They knew
that because of the nature of the GPL that improvements that were
distributed in the commercial environment could easily be folded back into
the standard version. Companies are not permitted to proprietarize
Samba, so the noncommercial users, and even other commercial users are
safe in the knowledge that the software freedom ensured by GPL will remain
protected.
Commercial developers also work in concert with noncommercial
developers. Those two now-long-since graduated students continue to
contribute to Samba altruistically, but also get paid work doing it.
Priorities change when a client is in the mix, but all the code is
contributed back to the standard version. Meanwhile, many other
individuals have gotten involved noncommercially as developers,
because they want to ``cut their teeth on Free Software,'' or because
the problems interest them. When they get good at it, perhaps they
will move on to another project, or perhaps they will become
commercial developers of the software themselves.
No party is a threat to another in the GPL software scenario because
everyone is on equal ground. The GPL protects rights of the commercial
and noncommercial contributors and users equally. The GPL creates trust,
because it is a level playing field for all.
\subsection{Law Analogy}
In his introduction to Stallman's \emph{Free Software, Free Society},
Lawrence Lessig draws an interesting analogy between the law and Free
Software. He argues that the laws of a free society must be protected
much like the GPL protects software. So that I might do true justice to
Lessig's argument, I quote it verbatim:
\begin{quotation}
A ``free society'' is regulated by law. But there are limits that any free
society places on this regulation through law: No society that kept its
laws secret could ever be called free. No government that hid its
regulations from the regulated could ever stand in our tradition. Law
controls. But it does so justly only when visibly. And law is visible
only when its terms are knowable and controllable by those it regulates,
or by the agents of those it regulates (lawyers, legislatures).
This condition on law extends beyond the work of a legislature. Think
about the practice of law in American courts. Lawyers are hired by their
clients to advance their clients' interests. Sometimes that interest is
advanced through litigation. In the course of this litigation, lawyers
write briefs. These briefs in turn affect opinions written by judges.
These opinions decide who wins a particular case, or whether a certain law
can stand consistently with a constitution.
All the material in this process is free in the sense that Stallman means.
Legal briefs are open and free for others to use. The arguments are
transparent (which is different from saying they are good), and the
reasoning can be taken without the permission of the original lawyers.
The opinions they produce can be quoted in later briefs. They can be
copied and integrated into another brief or opinion. The ``source code''
for American law is by design, and by principle, open and free for anyone
to take. And take lawyers do---for it is a measure of a great brief that
it achieves its creativity through the reuse of what happened before. The
source is free; creativity and an economy is built upon it.
This economy of free code (and here I mean free legal code) doesn't starve
lawyers. Law firms have enough incentive to produce great briefs even
though the stuff they build can be taken and copied by anyone else. The
lawyer is a craftsman; his or her product is public. Yet the crafting is
not charity. Lawyers get paid; the public doesn't demand such work
without price. Instead this economy flourishes, with later work added to
the earlier.
We could imagine a legal practice that was different --- briefs and
arguments that were kept secret; rulings that announced a result but not
the reasoning. Laws that were kept by the police but published to no one
else. Regulation that operated without explaining its rule.
We could imagine this society, but we could not imagine calling it
``free.'' Whether or not the incentives in such a society would be better
or more efficiently allocated, such a society could not be known as free.
The ideals of freedom, of life within a free society, demand more than
efficient application. Instead, openness and transparency are the
constraints within which a legal system gets built, not options to be
added if convenient to the leaders. Life governed by software code should
be no less.
Code writing is not litigation. It is better, richer, more
productive. But the law is an obvious instance of how creativity and
incentives do not depend upon perfect control over the products
created. Like jazz, or novels, or architecture, the law gets built
upon the work that went before. This adding and changing is what
creativity always is. And a free society is one that assures that its
most important resources remain free in just this sense.\footnote{This
quotation is Copyright \copyright{} 2002, Lawrence Lessig. It is
licensed under the terms of
\href{http://creativecommons.org/licenses/by/1.0/}{the ``Attribution
License'' version 1.0} or any later version as published by Creative
Commons.}
\end{quotation}
In essence, lawyers are paid to service the shared commons of legal
infrastructure. Few citizens defend themselves in court or write their
own briefs (even though they are legally permitted to do so) because
everyone would prefer to have an expert do that job.
The Free Software economy is a market ripe for experts. It
functions similarly to other well established professional fields like the
law. The GPL, in turn, serves as the legal scaffolding that permits the
creation of this vibrant commercial and noncommercial Free Software
economy.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\chapter{A Tale of Two Copyleft Licenses}
\label{tale-of-two-copylefts}
While determining the proper methodology and criteria to yield an accurate
count remains difficult, the GPL is generally considered one of the most
widely used Free Software licenses. For most of its history --- for 16 years
from June 1991 to June 2007 --- there was really only one version of the GPL,
version 2.
However, the GPL had both earlier versions before version 2, and, more well
known, a revision to version 3.
\section{Historical Motivations for the General Public License}
The earliest license to grant software freedom was likely the Berkeley
Software Distribution (``BSD'') license. This license is typical of what are
often called lax, highly permissive licenses. Not unlike software in the
public domain, these non-copyleft licenses (usually) grant software freedom
to users, but they do not go to any effort to uphold that software freedom
for users. The so-called ``downstream'' (those who receive the software and
then build new things based on that software) can restrict the software and
distribute further.
The GNU's Not Unix (``GNU'') project, which Richard M.~Stallman (``RMS'')
founded in 1984 to make a complete Unix-compatible operating system
implementation that assured software freedom for all. However, RMS saw that
using a license that gave but did not assure software freedom would be
counter to the goals of the GNU project. RMS invented ``copyleft'' as an
answer to that problem, and began using various copyleft licenses for the
early GNU project programs\footnote{RMS writes more fully about this topic in
his essay entitled simply
\href{http://www.gnu.org/gnu/thegnuproject.html}{\textit{The GNU Project}}.
For those who want to hear the story in his own voice,
\href{http://audio-video.gnu.org/audio/}{speech recordings} of his talk,
\textit{The Free Software Movement and the GNU/Linux Operating System}
are also widely available}.
\section{Proto-GPLs And Their Impact}
The earliest copyleft licenses were specific to various GNU programs. For
example, \href{http://www.free-soft.org/gpl_history/emacs_gpl.html}{The Emacs
General Public License} was likely the first copyleft license ever
published. Interesting to note that even this earliest copyleft license
contains a version of the well-known GPL copyleft clause:
\begin{quotation}
You may modify your copy or copies of GNU Emacs \ldots provided that you also
\ldots cause the whole of any work that you distribute or publish, that in
whole or in part contains or is a derivative of GNU Emacs or any part
thereof, to be licensed at no charge to all third parties on terms identical
to those contained in this License Agreement.
\end{quotation}
This simply stated clause is the fundamental innovation of copyleft.
Specifically, copyleft \textit{uses} the copyright holders' controls on
permission to modify the work to add a conditional requirement. Namely,
downstream users may only have permission to modify the work if they pass
along the same permissions on the modified version that came originally to
them.
These original program-specific proto-GPLs give an interesting window into
the central ideas and development of copyleft. In particular, reviewing them
shows how the text of the GPL we know has evolved to address more of the
issues discussed earlier in \S~\ref{software-and-non-copyright}.
\section{The GNU General Public License, Version 1}
\label{GPLv1}
In January 1989, the FSF announced that the GPL had been converted into a
``subroutine'' that could be reused not just for all FSF-copyrighted
programs, but also by anyone else. As the FSF claimed in its announcement of
the GPLv1\footnote{The announcement of GPLv1 was published in the
\href{http://www.gnu.org/bulletins/bull6.html\#SEC8}{GNU'S Bulletin, vol 1,
number 6 dated January 1989}. (Thanks very much to Andy Tai for his
\href{http://www.free-soft.org/gpl_history/}{consolidation of research on
the history of the pre-v1 GPL's}.)}:
\begin{quotation}
To make it easier to copyleft programs, we have been improving on the
legalbol architecture of the General Public License to produce a new version
that serves as a general-purpose subroutine: it can apply to any program
without modification, no matter who is publishing it.
\end{quotation}
This, like many inventive ideas, seems somewhat obvious in retrospect. But,
the FSF had some bright people and access to good lawyers when it started.
It took almost five years from the first copyleft licenses to get to a
generalized, reusable GPLv1. In the context and mindset of the 1980s, this
is not surprising. The idea of reusable licensing infrastructure was not
only uncommon, it was virtually nonexistent! Even the early BSD licenses
were simply copied and rewritten slightly for each new use\footnote{It
remains an interesting accident of history that the early BSD problematic
``advertising clause'' (discussion of which is somewhat beyond the scope of
this tutorial) lives on into current day, simply because while the
University of California at Berkeley gave unilateral permission to remove
the clause from \textit{its} copyrighted works, others who adapted the BSD
license with their own names in place of UC-Berkeley's never have.}. The
GPLv1's innovation of reuable licensing infrastructure, an obvious fact
today, was indeed a novel invention for its day\footnote{We're all just
grateful that the FSF also opposes business method patents, since the FSF's
patent on a ``method for reusable licensing infrastructure'' would have
not expired until 2006!}.
\section{The GNU General Public License, Version 2}
The GPLv2 was released two and a half years after GPLv1, and over the
following sixteen years, it became the standard for copyleft licensing until
the release of GPLv3 in 2007 (discussed in more detail in the next section).
While this tutorial does not discuss the terms of GPLv1 in detail, it is
worth noting below the three key changes that GPLv2 brought:
\begin{itemize}
\item Software patents and their danger are explicitly mentioned, inspiring
(in part) the addition of GPLv2~\S\S5--7. (These sections are discussed in
detail in \S~\ref{GPLv2s5}, \S~\ref{GPLv2s6} and \S~\ref{GPLv2s7} of this
tutorial.)
\item GPLv2~\S2's copyleft terms are expanded to more explicitly discuss the
issue of combined works. (GPLv2~\S2 is discussed in detail in
\S~\ref{GPLv2s2} in this tutorial).
\item GPLv2~\S3 includes more detailed requirements, including the phrase
``the scripts used to control compilation and installation of the
executable'', which is a central component of current GPLv2 enforcement
. (GPLv2~\S3 is discussed in detail in
\S~\ref{GPLv2s3} in this tutorial).
\end{itemize}
The next chapter discusses GPLv2 in full detail, and readers who wish to dive
into the section-by-section discussion of the GPL should jump ahead now to
that chapter. However, the most interesting fact to note here is how GPLv2
was published with little fanfare and limited commentary. This contrasts
greatly with the creation of GPLv3.
\section{The GNU General Public License, Version 3}
RMS began drafting GPLv2.2 in mid-2002, and FSF ran a few discussion groups
during that era about new text of that license. However, rampant violations
of the GPL required more immediate attention of FSF's licensing staff, and as
such, much of the early 2000's was spent doing GPL enforcement
work\footnote{More on GPL enforcement is discussed in \tutorialpartsplit{a
companion tutorial, \textit{A Practical Guide to GPL
Compliance}}{Part~\ref{gpl-compliance-guide} of this tutorial}.}. In
2006, FSF began in earnest drafting work for GPLv3.
The GPLv3 process began in earnest in January 2006. It became clear that
many provisions of the GPL could benefit from modification to fit new
circumstances and to reflect what the entire community learned from
experience with version 2. Given the scale of revision it seems proper to
approach the work through public discussion in a transparent and accessible
manner.
The GPLv3 process continued through June 2007, culminating in publication of
GPLv3 and LGPLv3 on 29 June 2007, AGPLv3 on 19 November 2007, and the GCC
Runtime Library Exception on 27 January 2009.
All told, four discussion drafts of GPLv3, two discussion drafts of LGPLv3
and two discussion drafts of AGPLv3 were published and discussed.
Ultimately, FSF remained the final arbiter and publisher of the licenses, and
RMS himself their primary author, but input was sought from many parties, and
these licenses do admittedly look and read more like legislation as a result.
Nevertheless, all of the ``v3'' group are substantially better and improved
licenses.
GPLv3 and its terms are discussed in detail in Chapter\~ref{GPLv3}.
\section{The Innovation of Optional ``Or Any Later'' Version}
An interesting fact of all GPL licenses is that the are ultimate multiple
choices for use of the license. The FSF is the primary steward of GPL (as
discussed later in \S~\ref{GPLv2s9} and \S~\ref{GPLv2s14}). However, those
who wish to license works under GPL are not required to automatically accept
changes made by the FSF for their own copyrighted works.
Each licensor may chose three different methods of licensing, as follows:
\begin{itemize}
\item explicitly name a single version of GPL for their work (usually
indicated in shorthand by saying the license is ``GPLv$X$-only''), or
\item name no version of the GPL, thus they allow their downstream recipients
to select any version of the GPL they chose (usually indicated in shorthand
by saying the license is simply ``GPL''), or
\item name a specific version of GPL and give downstream recipients the
option to chose that version ``or any later version as published by the
FSF'' (usually indicated by saying the license is
``GPLv$X$-or-later'')\footnote{The shorthand of ``GPL$X+$'' is also popular
for this situation. The authors of this tutorial prefer ``-or-later''
syntax, because it (a) mirrors the words ``or'' and ``later from the
licensing statement, (b) the $X+$ doesn't make it abundantly clear that
$X$ is clearly included as a license option and (c) the $+$ symbol has
other uses in computing (such as with regular expressions) that mean
something different.}
\end{itemize}
\label{license-compatibility-first-mentioned}
Oddly, this flexibility has received (in the opinion of the authors, undue)
criticism, primarily because of the complex and oft-debated notion of
``license compatibility'' (which is explained in detail in
\S~\ref{license-compatibility}). Copyleft licenses are generally
incompatible with each other, because the details of how they implement
copyleft differs. Specifically, copyleft works only because of its
requirement that downstream licensors use the \textit{same} license for
combined and modified works. As such, software licensed under the terms of
``GPLv2-only'' cannot be combined with works licensed ``GPLv3-or-later''.
This is admittedly a frustrating outcome.
Other copyleft licenses that appeared after GPL, such
as the Creative Commons ``Share Alike'' licenses, the Eclipse Public License
and the Mozilla Public License \textbf{require} all copyright holders chosing
to use any version of those licenses to automatically accept and relicense
their copyrighted works under new versions. Of course ,Creative Commons, the
Eclipse Foundation, and the Mozilla Foundation (like the FSF) have generally
served as excellent stewards of their licenses. Copyright holders using
those licenses seems to find it acceptable that to fully delegate all future
licensing decisions for their copyrights to these organizations without a
second thought.
However, note that FSF gives herein the control of copyright holders to
decide whether or not to implicitly trust the FSF in its work of drafting
future GPL versions. The FSF, for its part, does encourage copyright holders
to chose by default ``GPLv$X$-or-later'' (where $X$ is the most recent
version of the GPL published by the FSF). However, the FSF \textbf{does not
mandate} that a choice to use any GPL requires a copyright holder ceding
its authority for future licensing decisions to the FSF. In fact, the FSF
considered this possibility for GPLv3 and chose not to do so, instead opting
for the third-party steward designation clause discussed in
Section~\ref{GPLv3s14}.
\section{Complexities of Two Simultaneously Popular Copylefts}
Obviously most GPL advocates would prefer widespread migration to GPLv3, and
many newly formed projects who seek a copyleft license tend to choose a
GPLv3-based license. However, many existing copylefted projects continue
with GPLv2-only or GPLv2-or-later as their default license.
While GPLv3 introduces many improvements --- many of which were designed to
increase adoption by for-profit companies --- GPLv2 remains a widely used and
extremely popular license. The GPLv2 is, no doubt, a good and useful
license.
However, unlike GPLv1, which (as pointed out in \S~\ref{GPLv1}), which is
completely out of use by the mid-1990s. However, unlike GPLv1 before it,
GPLv2 remains a integral part of the copyleft licensing infrastructure for
some time to come. As such, those who seek to have expertise in current
topics of copyleft licensing need to study both the GPLv2 and GPLv3 family of
licenses.
Furthermore, GPLv3 can is more easily understood by first studying GPLv2.
This is not only because of their chronological order, but also because much
of the discussion material available for GPLv3 tends to talk about GPLv3 in
contrast to GPLv2. As such, a strong understanding of GPLv2 helps in
understanding most of the third-party material found regarding GPLv3. Thus,
the following chapter begins a deep discussion of GPLv2.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\chapter{Running Software and Verbatim Copying}
\label{run-and-verbatim}
This chapter begins the deep discussion of the details of the terms of
GPLv2\@. In this chapter, we consider the first two sections: GPLv2 \S\S
0--2. These are the straightforward sections of the GPL that define the
simplest rights that the user receives.
\section{GPLv2~\S0: Freedom to Run}
\label{GPLv2s0}
GPLv2~\S0, the opening section of GPLv2, sets forth that the copyright law governs
the work. It specifically points out that it is the ``copyright
holder'' who decides if a work is licensed under its terms and explains
how the copyright holder might indicate this fact.
A bit more subtly, GPLv2~\S0 makes an inference that copyright law is the only
system that can restrict the software. Specifically, it states:
\begin{quote}
Activities other than copying, distribution and modification are not
covered by this License; they are outside its scope.
\end{quote}
In essence, the license governs \emph{only} those activities, and all other
activities are unrestricted, provided that no other agreements trump GPLv2
(which they cannot; see Sections~\ref{GPLv2s6} and~\ref{GPLv2s7}). This is
very important, because the Free Software community heavily supports
users' rights to ``fair use'' and ``unregulated use'' of copyrighted
material. GPLv2 asserts through this clause that it supports users' rights
to fair and unregulated uses.
Fair use (called ``fair dealing'' in some jurisdictions) of copyrighted
material is an established legal doctrine that permits certain activities
regardless of whether copyright law would other restrict those activities.
Discussion of the various types of fair use activity are beyond the scope of
this tutorial. However, one important example of fair use is the right to
quote portions of the text in larger work so as to criticize or suggest
changes. This fair use rights is commonly used on mailing lists when
discussing potential improvements or changes to Free Software.
Fair use is a doctrine established by the courts or by statute. By
contrast, unregulated uses are those that are not covered by the statue
nor determined by a court to be covered, but are common and enjoyed by
many users. An example of unregulated use is reading a printout of the
program's source code like an instruction book for the purpose of learning
how to be a better programmer. The right to read something that you have
access is and should remain unregulated and unrestricted.
\medskip
Thus, the GPLv2 protects users fair and unregulated use rights precisely by
not attempting to cover them. Furthermore, the GPLv2 ensures the freedom
to run specifically by stating the following:
\begin{quote}
''The act of running the Program is not restricted.''
\end{quote}
Thus, users are explicitly given the freedom to run by GPLv2~\S0.
\medskip
The bulk of GPLv2~\S0 not yet discussed gives definitions for other terms used
throughout. The only one worth discussing in detail is ``work based on
the Program''. The reason this definition is particularly interesting is
not for the definition itself, which is rather straightforward, but
because it clears up a common misconception about the GPL\@.
The GPL is often mistakenly criticized because it fails to give a
definition of ``derivative work''. In fact, it would be incorrect and
problematic if the GPL attempted to define this. A copyright license, in
fact, has no control over what may or may not be a derivative work. This
matter is left up to copyright law and the courts --- not the licenses that utilize it.
It is certainly true that copyright law as a whole does not propose clear
and straightforward guidelines for what is and is not a derivative
software work under copyright law. However, no copyright license --- not
even the GNU GPL --- can be blamed for this. Legislators and court
opinions must give us guidance to decide the border cases.
\section{GPLv2~\S1: Verbatim Copying}
\label{GPLv2s1}
GPLv2~\S1 covers the matter of redistributing the source code of a program
exactly as it was received. This section is quite straightforward.
However, there are a few details worth noting here.
The phrase ``in any medium'' is important. This, for example, gives the
freedom to publish a book that is the printed copy of the program's source
code. It also allows for changes in the medium of distribution. Some
vendors may ship Free Software on a CD, but others may place it right on
the hard drive of a pre-installed computer. Any such redistribution media
is allowed.
Preservation of copyright notice and license notifications are mentioned
specifically in GPLv2~\S1. These are in some ways the most important part of
the redistribution, which is why they are mentioned by name. GPL
always strives to make it abundantly clear to anyone who receives the
software what its license is. The goal is to make sure users know their
rights and freedoms under GPL, and to leave no reason that users might be
surprised the software is GPL'd. Thus
throughout the GPL, there are specific references to the importance of
notifying others down the distribution chain that they have rights under
GPL.
Also mentioned by name is the warranty disclaimer. Most people today do
not believe that software comes with any warranty. Notwithstanding the
\href{http://mlis.state.md.us/2000rs/billfile/hb0019.htm}{Maryland's} and \href{http://leg1.state.va.us/cgi-bin/legp504.exe?001+ful+SB372ER}{Virginia's} UCITA bills, there are few or no implied warranties with software.
However, just to be on the safe side, GPL clearly disclaims them, and the
GPL requires redistributors to keep the disclaimer very visible. (See
Sections~\ref{GPLv2s11} and~\ref{GPLv2s12} of this tutorial for more on GPL's
warranty disclaimers.)
Note finally that GPLv2~\S1 creates groundwork for the important defense of
commercial freedom. GPLv2~\S1 clearly states that in the case of verbatim
copies, one may make money. Redistributors are fully permitted to charge
for the redistribution of copies of Free Software. In addition, they may
provide the warranty protection that the GPL disclaims as an additional
service for a fee. (See Section~\ref{Business Models} for more discussion
on making a profit from Free Software redistribution.)
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\chapter{Derivative Works: Statute and Case Law}
We digress for this chapter from our discussion of GPL's exact text to
consider the matter of derivative works --- a concept that we must
understand fully before considering GPLv2~\S\S2--3\@. GPL, and Free
Software licensing in general, relies critically on the concept of
``derivative work'' since software that is ``independent,'' (i.e., not
``derivative'') of Free Software need not abide by the terms of the
applicable Free Software license. As much is required by \S~106 of the
Copyright Act, 17 U.S.C. \S~106 (2002), and admitted by Free Software
licenses, such as the GPL, which (as we have seen) states in GPLv2~\S0 that ``a
`work based on the Program' means either the Program or any derivative
work under copyright law.'' It is being a derivative work of Free Software
that triggers the necessity to comply with the terms of the Free Software
license under which the original work is distributed. Therefore, one is
left to ask, just what is a ``derivative work''? The answer to that
question differs depending on which court is being asked.
The analysis in this chapter sets forth the differing definitions of
derivative work by the circuit courts. The broadest and most
established definition of derivative work for software is the
abstraction, filtration, and comparison test (``the AFC test'') as
created and developed by the Second Circuit. Some circuits, including
the Ninth Circuit and the First Circuit, have either adopted narrower
versions of the AFC test or have expressly rejected the AFC test in
favor of a narrower standard. Further, several other circuits have yet
to adopt any definition of derivative work for software.
As an introductory matter, it is important to note that literal copying of
a significant portion of source code is not always sufficient to establish
that a second work is a derivative work of an original
program. Conversely, a second work can be a derivative work of an original
program even though absolutely no copying of the literal source code of
the original program has been made. This is the case because copyright
protection does not always extend to all portions of a program's code,
while, at the same time, it can extend beyond the literal code of a
program to its non-literal aspects, such as its architecture, structure,
sequence, organization, operational modules, and computer-user interface.
\section{The Copyright Act}
The copyright act is of little, if any, help in determining the definition
of a derivative work of software. However, the applicable provisions do
provide some, albeit quite cursory, guidance. Section 101 of the Copyright
Act sets forth the following definitions:
\begin{quotation}
A ``computer program'' is a set of statements or instructions to be used
directly or indirectly in a computer in order to bring about a certain
result.
A ``derivative work'' is a work based upon one or more preexisting works,
such as a translation, musical arrangement, dramatization,
fictionalization, motion picture version, sound recording, art
reproduction, abridgment, condensation, or any other form in which a work
may be recast, transformed, or adapted. A work consisting of editorial
revisions, annotations, elaborations, or other modifications which, as a
whole, represent an original work of authorship, is a ``derivative work.''
\end{quotation}
These are the only provisions in the Copyright Act relevant to the
determination of what constitutes a derivative work of a computer
program. Another provision of the Copyright Act that is also relevant to
the definition of derivative work is \S~102(b), which reads as follows:
\begin{quotation}
In no case does copyright protection for an original work of authorship
extend to any idea, procedure, process, system, method of operation,
concept, principle, or discovery, regardless of the form in which it is
described, explained, illustrated, or embodied in such work.
\end{quotation}
Therefore, before a court can ask whether one program is a derivative work
of another program, it must be careful not to extend copyright protection
to any ideas, procedures, processes, systems, methods of operation,
concepts, principles, or discoveries contained in the original program. It
is the implementation of this requirement to ``strip out'' unprotectable
elements that serves as the most frequent issue over which courts
disagree.
\section{Abstraction, Filtration, Comparison Test}
As mentioned above, the AFC test for determining whether a computer
program is a derivative work of an earlier program was created by the
Second Circuit and has since been adopted in the Fifth, Tenth, and
Eleventh Circuits. Computer Associates Intl., Inc. v. Altai, Inc., 982
F.2d 693 (2nd Cir. 1992); Engineering Dynamics, Inc. v. Structural
Software, Inc., 26 F.3d 1335 (5th Cir. 1994); Kepner-Tregoe,
Inc. v. Leadership Software, Inc., 12 F.3d 527 (5th Cir. 1994); Gates
Rubber Co. v. Bando Chem. Indust., Ltd., 9 F.3d 823 (10th Cir. 1993);
Mitel, Inc. v. Iqtel, Inc., 124 F.3d 1366 (10th Cir. 1997); Bateman
v. Mnemonics, Inc., 79 F.3d 1532 (11th Cir. 1996); and, Mitek Holdings,
Inc. v. Arce Engineering Co., Inc., 89 F.3d 1548 (11th Cir. 1996).
Under the AFC test, a court first abstracts from the original program its
constituent structural parts. Then, the court filters from those
structural parts all unprotectable portions, including incorporated ideas,
expression that is necessarily incidental to those ideas, and elements
that are taken from the public domain. Finally, the court compares any and
all remaining kernels of creative expression to the structure of the
second program to determine whether the software programs at issue are
substantially similar so as to warrant a finding that one is the
derivative work of the other.
Often, the courts that apply the AFC test will perform a quick initial
comparison between the entirety of the two programs at issue in order to
help determine whether one is a derivative work of the other. Such a
holistic comparison, although not a substitute for the full application of
the AFC test, sometimes reveals a pattern of copying that is not otherwise
obvious from the application of the AFC test when, as discussed below,
only certain components of the original program are compared to the second
program. If such a pattern is revealed by the quick initial comparison,
the court is more likely to conclude that the second work is indeed a
derivative of the original.
\subsection{Abstraction}
The first step courts perform under the AFC test is separation of the
work's ideas from its expression. In a process akin to reverse
engineering, the courts dissect the original program to isolate each level
of abstraction contained within it. Courts have stated that the
abstractions step is particularly well suited for computer programs
because it breaks down software in a way that mirrors the way it is
typically created. However, the courts have also indicated that this step
of the AFC test requires substantial guidance from experts, because it is
extremely fact and situation specific.
By way of example, one set of abstraction levels is, in descending order
of generality, as follows: the main purpose, system architecture, abstract
data types, algorithms and data structures, source code, and object
code. As this set of abstraction levels shows, during the abstraction step
of the AFC test, the literal elements of the computer program, namely the
source and object code, are defined as particular levels of
abstraction. Further, the source and object code elements of a program are
not the only elements capable of forming the basis for a finding that a
second work is a derivative of the program. In some cases, in order to
avoid a lengthy factual inquiry by the court, the owner of the copyright in
the original work will submit its own list of what it believes to be the
protected elements of the original program. In those situations, the court
will forgo performing its own abstraction, and proceed to the second step of
the AFC test.
\subsection{Filtration}
The most difficult and controversial part of the AFC test is the second
step, which entails the filtration of protectable expression contained in
the original program from any unprotectable elements nestled therein. In
determining which elements of a program are unprotectable, courts employ a
myriad of rules and procedures to sift from a program all the portions
that are not eligible for copyright protection.
First, as set forth in \S~102(b) of the Copyright Act, any and all ideas
embodied in the program are to be denied copyright protection. However,
implementing this rule is not as easy as it first appears. The courts
readily recognize the intrinsic difficulty in distinguishing between ideas
and expression and that, given the varying nature of computer programs,
doing so will be done on an ad hoc basis. The first step of the AFC test,
the abstraction, exists precisely to assist in this endeavor by helping
the court separate out all the individual elements of the program so that
they can be independently analyzed for their expressive nature.
A second rule applied by the courts in performing the filtration step of
the AFC test is the doctrine of merger, which denies copyright protection
to expression necessarily incidental to the idea being expressed. The
reasoning behind this doctrine is that when there is only one way to
express an idea, the idea and the expression merge, meaning that the
expression cannot receive copyright protection due to the bar on copyright
protection extending to ideas. In applying this doctrine, a court will ask
whether the program's use of particular code or structure is necessary for
the efficient implementation of a certain function or process. If so, then
that particular code or structure is not protected by copyright and, as a
result, it is filtered away from the remaining protectable expression.
A third rule applied by the courts in performing the filtration step of
the AFC test is the doctrine of scenes a faire, which denies copyright
protection to elements of a computer program that are dictated by external
factors. Such external factors can include:
\begin{itemize}
\item The mechanical
specifications of the computer on which a particular program is intended
to operate
\item Compatibility requirements of other programs with which a
program is designed to operate in conjunction
\item Computer manufacturers'
design standards
\item Demands of the industry being serviced, and
widely accepted programming practices within the computer industry
\end{itemize}
Any code or structure of a program that was shaped predominantly in
response to these factors is filtered out and not protected by
copyright. Lastly, elements of a computer program are also to be filtered
out if they were taken from the public domain or fail to have sufficient
originality to merit copyright protection.
Portions of the source or object code of a computer program are rarely
filtered out as unprotectable elements. However, some distinct parts of
source and object code have been found unprotectable. For example,
constant s, the invariable integers comprising part of formulas used to
perform calculations in a program, are unprotectable. Further, although
common errors found in two programs can provide strong evidence of
copying, they are not afforded any copyright protection over and above the
protection given to the expression containing them.
\subsection{Comparison}
The third and final step of the AFC test entails a comparison of the
original program's remaining protectable expression to a second
program. The issue will be whether any of the protected expression is
copied in the second program and, if so, what relative importance the
copied portion has with respect to the original program overall. The
ultimate inquiry is whether there is ``substantial'' similarity between
the protected elements of the original program and the potentially
derivative work. The courts admit that this process is primarily
qualitative rather than quantitative and is performed on a case-by-case
basis. In essence, the comparison is an ad hoc determination of whether
the protectable elements of the original program that are contained in the
second work are significant or important parts of the original program. If
so, then the second work is a derivative work of the first. If, however,
the amount of protectable elements copied in the second work are so small
as to be de minimis, then the second work is not a derivative work of the
original.
\section{Analytic Dissection Test}
The Ninth Circuit has adopted the analytic dissection test to determine
whether one program is a derivative work of another. Apple Computer,
Inc. v. Microsoft Corp., 35 F.3d 1435 (9th Cir. 1994). The analytic
dissection test first considers whether there are substantial similarities
in both the ideas and expressions of the two works at issue. Once the
similar features are identified, analytic dissection is used to determine
whether any of those similar features are protected by copyright. This
step is the same as the filtration step in the AFC test. After identifying
the copyrightable similar features of the works, the court then decides
whether those features are entitled to ``broad'' or ``thin''
protection. ``Thin'' protection is given to non-copyrightable facts or
ideas that are combined in a way that affords copyright protection only
from their alignment and presentation, while ``broad'' protection is given
to copyrightable expression itself. Depending on the degree of protection
afforded, the court then sets the appropriate standard for a subjective
comparison of the works to determine whether, as a whole, they are
sufficiently similar to support a finding that one is a derivative work of
the other. ``Thin'' protection requires the second work be virtually
identical in order to be held a derivative work of an original, while
``broad'' protection requires only a ``substantial similarity.''
\section{No Protection for ``Methods of Operation''}
The First Circuit has taken the position that the AFC test is inapplicable
when the works in question relate to unprotectable elements set forth in
\S~102(b). Their approach results in a much narrower definition
of derivative work for software in comparison to other circuits. Specifically,
the
First Circuit holds that ``method of operation,'' as used in \S~102(b) of
the Copyright Act, refers to the means by which users operate
computers. Lotus Development Corp. v. Borland Int’l., Inc., 49 F.3d 807
(1st Cir. 1995). In Lotus, the court held that a menu command
hierarchy for a computer program was uncopyrightable because it did not
merely explain and present the program’s functional capabilities to the
user, but also served as a method by which the program was operated and
controlled. As a result, under the First Circuit’s test, literal copying
of a menu command hierarchy, or any other ``method of operation,'' cannot
form the basis for a determination that one work is a derivative of
another. As a result, courts in the First Circuit that apply the AFC test
do so only after applying a broad interpretation of \S~102(b) to filter out
unprotected elements. E.g., Real View, LLC v. 20-20 Technologies, Inc.,
683 F. Supp.2d 147, 154 (D. Mass. 2010).
\section{No Test Yet Adopted}
Several circuits, most notably the Fourth and Seventh, have yet to
declare their definition of derivative work and whether or not the
AFC, Analytic Dissection, or some other test best fits their
interpretation of copyright law. Therefore, uncertainty exists with
respect to determining the extent to which a software program is a
derivative work of another in those circuits. However, one may presume
that they would give deference to the AFC test since it is by far the
majority rule amongst those circuits that have a standard for defining
a software derivative work.
\section{Cases Applying Software Derivative Work Analysis}
In the preeminent case regarding the definition of a derivative work for
software, Computer Associates v. Altai, the plaintiff alleged that its
program, Adapter, which was used to handle the differences in operating
system calls and services, was infringed by the defendant's competitive
program, Oscar. About 30\% of Oscar was literally the same code as
that in Adapter. After the suit began, the defendant rewrote those
portions of Oscar that contained Adapter code in order to produce a new
version of Oscar that was functionally competitive with Adapter, without
have any literal copies of its code. Feeling slighted still, the
plaintiff alleged that even the second version of Oscar, despite having no
literally copied code, also infringed its copyrights. In addressing that
question, the Second Circuit promulgated the AFC test.
In abstracting the various levels of the program, the court noted a
similarity between the two programs' parameter lists and macros. However,
following the filtration step of the AFC test, only a handful of the lists
and macros were protectable under copyright law because they were either
in the public domain or required by functional demands on the
program. With respect to the handful of parameter lists and macros that
did qualify for copyright protection, after performing the comparison step
of the AFC test, it was reasonable for the district court to conclude that
they did not warrant a finding of infringement given their relatively minor
contribution to the program as a whole. Likewise, the similarity between
the organizational charts of the two programs was not substantial enough
to support a finding of infringement because they were too simple and
obvious to contain any original expression.
In the case of Oracle America v. Google, 872 F. Supp.2d 974 (N.D. Cal. 2012),
the Northern District of California District Court examined the question of
whether the application program interfaces (APIs) associated with the Java
programming language are entitled to copyright protection. While the
court expressly declined to rule whether all APIs are free to use without
license (872 F. Supp.2d 974 at 1002), the court held that the command
structure and taxonomy of the APIs were not protectable under copyright law.
Specifically, the court characterized the command structure and taxonomy as
both a ``method of operation'' (using an approach not dissimilar to the
First Circuit's analysis in Lotus) and a ``functional requirement for
compatability'' (using Sega v. Accolade, 977 F.2d 1510 (9th Cir. 1992) and
Sony Computer Ent. v. Connectix, 203 F.3d 596 (9th Cir. 2000) as analogies),
and thus unprotectable subject matter under \S~102(b).
Perhaps not surprisingly, there have been few other cases involving a highly
detailed software derivative work analysis. Most often, cases involve
clearer basis for decision, including frequent bad faith on the part of
the defendant or overaggressiveness on the part of the plaintiff.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\chapter{Modified Source and Binary Distribution}
\label{source-and-binary}
In this chapter, we discuss the two core sections that define the rights
and obligations for those who modify, improve, and/or redistribute GPL'd
software. These sections, GPLv2~\S\S2--3, define the central core rights and
requirements of GPLv2\@.
\section{GPLv2~\S2: Share and Share Alike}
\label{GPLv2s2}
For many, this is where the ``magic'' happens that defends software
freedom upon redistribution. GPLv2~\S2 is the only place in GPLv2
that governs the modification controls of copyright law. If users
modifies a GPLv2'd program, they must follow the terms of GPLv2~\S2 in making
those changes. Thus, this sections ensures that the body of GPL'd software, as it
continues and develops, remains Free as in freedom.
To achieve that goal, GPLv2~\S2 first sets forth that the rights of
redistribution of modified versions are the same as those for verbatim
copying, as presented in GPLv2~\S1. Therefore, the details of charging money,
keeping copyright notices intact, and other GPLv2~\S1 provisions are in tact
here as well. However, there are three additional requirements.
The first (GPLv2~\S2(a)) requires that modified files carry ``prominent
notices'' explaining what changes were made and the date of such
changes. This section does not prescribe some specific way of
marking changes nor does it control the process of how changes are made.
Primarily, GPLv2~\S2(a) seeks to ensure that those receiving modified
versions know the history of changes to the software. For some users,
it is important to know that they are using the standard version of
program, because while there are many advantages to using a fork,
there are a few disadvantages. Users should be informed about the
historical context of the software version they use, so that they can
make proper support choices. Finally, GPLv2~\S2(a) serves an academic
purpose --- ensuring that future developers can use a diachronic
approach to understand the software.
\medskip
The second requirement (GPLv2~\S2(b)) contains the four short lines that embody
the legal details of ``share and share alike''. These 46 words are
considered by some to be the most worthy of careful scrutiny because
GPLv2~\S2(b), and they
can be a source of great confusion when not properly understood.
In considering GPLv2~\S2(b), first note the qualifier: it \textit{only} applies to
derivative works that ``you distribute or publish''. Despite years of
education efforts on this matter, many still believe that modifiers
of GPL'd software \textit{must} to publish or otherwise
share their changes. On the contrary, GPLv2~\S2(b) {\bf does not apply if} the
changes are never distributed. Indeed, the freedom to make private,
personal, unshared changes to software for personal use only should be
protected and defended.\footnote{Most Free Software enthusiasts believe there is an {\bf
moral} obligation to redistribute changes that are generally useful,
and they often encourage companies and individuals to do so. However, there
is a clear distinction between what one {\bf ought} to do and what one
{\bf must} do.}
Next, we again encounter the same matter that appears in GPLv2~\S0, in the
following text:
\begin{quote}
``...that in whole or part contains or is derived from the Program or any part thereof.''
\end{quote}
Again, the GPL relies here on what the copyright law says is a derivative
work. If, under copyright law, the modified version ``contains or is
derived from'' the GPL'd software, then the requirements of GPLv2~\S2(b)
apply. The GPL invokes its control as a copyright license over the
modification of the work in combination with its control over distribution
of the work.
The final clause of GPLv2~\S2(b) describes what the licensee must do if she is
distributing or publishing a work that is deemed a derivative work under
copyright law --- namely, the following:
\begin{quote}
[The work must] be licensed as a whole at no charge to all third parties
under the terms of this License.
\end{quote}
That is probably the most tightly-packed phrase in all of the GPL\@.
Consider each subpart carefully.
The work ``as a whole'' is what is to be licensed. This is an important
point that GPLv2~\S2 spends an entire paragraph explaining; thus this phrase is
worthy of a lengthy discussion here. As a programmer modifies a software
program, she generates new copyrighted material --- fixing expressions of
ideas into the tangible medium of electronic file storage. That
programmer is indeed the copyright holder of those new changes. However,
those changes are part and parcel to the original work distributed to
the programmer under GPL\@. Thus, the license of the original work
affects the license of the new whole derivative work.
% {\cal I}
\newcommand{\gplusi}{$\mathcal{G\!\!+\!\!I}$}
\newcommand{\worki}{$\mathcal{I}$}
\newcommand{\workg}{$\mathcal{G}$}
\label{separate-and-independent}
It is certainly possible to take an existing independent work (called
\worki{}) and combine it with a GPL'd program (called \workg{}). The
license of \worki{}, when it is distributed as a separate and independent
work, remains the prerogative of the copyright holder of \worki{}.
However, when \worki{} is combined with \workg{}, it produces a new work
that is the combination of the two (called \gplusi{}). The copyright of
this combined work, \gplusi{}, is held by the original copyright
holder of each of the two works.
In this case, GPLv2~\S2 lays out the terms by which \gplusi{} may be
distributed and copied. By default, under copyright law, the copyright
holder of \worki{} would not have been permitted to distribute \gplusi{};
copyright law forbids it without the expressed permission of the copyright
holder of \workg{}. (Imagine, for a moment, if \workg{} were a proprietary
product --- would its copyright holders give you permission to create and distribute
\gplusi{} without paying them a hefty sum?) The license of \workg{}, the
GPL, states the options for the copyright holder of \worki{}
who may want to create and distribute \gplusi{}. GPL's pregranted
permission to create and distribute derivative works, provided the terms
of GPL are upheld, goes far above and beyond the permissions that one
would get with a typical work not covered by a copyleft license. (Thus, to
say that this restriction is any way unreasonable is simply ludicrous.)
\medskip
\label{GPLv2s2-at-no-charge}
The next phrase of note in GPLv2~\S2(b) is ``licensed \ldots at no charge.''
This phrase confuses many. The sloppy reader points out this as ``a
contradiction in GPL'' because (in their confused view) that clause of GPLv2~\S2 says that redistributors cannot
charge for modified versions of GPL'd software, but GPLv2~\S1 says that
they can. Avoid this confusion: the ``at no charge'' \textbf{does not} prohibit redistributors from
charging when performing the acts governed by copyright
law,\footnote{Recall that you could by default charge for any acts not
governed by copyright law, because the license controls are confined
by copyright.} but rather that they cannot charge a fee for the
\emph{license itself}. In other words, redistributors of (modified
and unmodified) GPL'd works may charge any amount they choose for
performing the modifications on contract or the act of transferring
the copy to the customer, but they may not charge a separate licensing
fee for the software.
GPLv2~\S2(b) further states that the software must ``be licensed \ldots to all
third parties.'' This too yields some confusion, and feeds the
misconception mentioned earlier --- that all modified versions must made
available to the public at large. However, the text here does not say
that. Instead, it says that the licensing under terms of the GPL must
extend to anyone who might, through the distribution chain, receive a copy
of the software. Distribution to all third parties is not mandated here,
but GPLv2~\S2(b) does require redistributors to license the derivative works in
a way that extends to all third parties who may ultimately receive a
copy of the software.
In summary, GPLv2\ 2(b) says what terms under which the third parties must
receive this no-charge license. Namely, they receive it ``under the terms
of this License'', the GPLv2. When an entity \emph{chooses} to redistribute
a derivative work of GPL'd software, the license of that whole
work must be GPL and only GPL\@. In this manner, GPLv2~\S2(b) dovetails nicely
with GPLv2~\S6 (as discussed in Section~\ref{GPLv2s6} of this tutorial).
\medskip
The final paragraph of GPLv2~\S2 is worth special mention. It is possible and
quite common to aggregate various software programs together on one
distribution medium. Computer manufacturers do this when they ship a
pre-installed hard drive, and GNU/Linux distribution vendors do this to
give a one-stop CD or URL for a complete operating system with necessary
applications. The GPL very clearly permits such ``mere aggregation'' with
programs under any license. Despite what you hear from its critics, the
GPL is nothing like a virus, not only because the GPL is good for you and
a virus is bad for you, but also because simple contact with a GPL'd
code-base does not impact the license of other programs. A programmer must
expended actual effort to cause a work to fall under the terms
of the GPL. Redistributors are always welcome to simply ship GPL'd
software alongside proprietary software or other unrelated Free Software,
as long as the terms of GPL are adhered to for those packages that are
truly GPL'd.
\section{GPLv2~\S3: Producing Binaries}
\label{GPLv2s3}
Software is a strange beast when compared to other copyrightable works.
It is currently impossible to make a film or a book that can be truly
obscured. Ultimately, the full text of a novel, even one written by
William Faulkner, must presented to the reader as words in some
human-readable language so that they can enjoy the work. A film, even one
directed by David Lynch, must be perceptible by human eyes and ears to
have any value.
Software is not so. While the source code --- the human-readable
representation of software is of keen interest to programmers -- users and
programmers alike cannot make the proper use of software in that
human-readable form. Binary code --- the ones and zeros that the computer
can understand --- must be predicable and attainable for the software to
be fully useful. Without the binaries, be they in object or executable
form, the software serves only the didactic purposes of computer science.
Under copyright law, binary representations of the software are simply
derivative works of the source code. Applying a systematic process (i.e.,
``compilation''\footnote{``Compilation'' in this context refers to the
automated computing process of converting source code into binaries. It
has absolutely nothing to do with the term ``compilation'' in copyright statues.}) to a work of source code yields binary code. The binary
code is now a new work of expression fixed in the tangible medium of
electronic file storage.
Therefore, for GPL'd software to be useful, the GPL, since it governs the
rules for creation of derivative works, must grant permission for the
generation of binaries. Furthermore, notwithstanding the relative
popularity of source-based GNU/Linux distributions like Gentoo, users find
it extremely convenient to receive distribution of binary software. Such
distribution is the redistribution of derivative works of the software's
source code. GPLv2~\S3 addresses the matter of creation and distribution of
binary versions.
Under GPLv2~\S3, binary versions may be created and distributed under the
terms of GPLv2~\S1--2, so all the material previously discussed applies
here. However, GPLv2~\S3 must go a bit further. Access to the software's
source code is an incontestable prerequisite for the exercise of the
fundamental freedoms to modify and improve the software. Making even
the most trivial changes to a software program at the binary level is
effectively impossible. GPLv2~\S3 must ensure that the binaries are never
distributed without the source code, so that these freedoms are passed
through the distribution chain.
GPLv2~\S3 permits distribution of binaries, and then offers three options for
distribution of source code along with binaries. The most common and the
least complicated is the option given under GPLv2~\S3(a).
GPLv2~\S3(a) offers the option to directly accompany the source code alongside
the distribution of the binaries. This is by far the most convenient
option for most distributors, because it means that the source-code
provision obligations are fully completed at the time of binary
distribution (more on that later).
Under GPLv2~\S3(a), the source code provided must be the ``corresponding source
code.'' Here ``corresponding'' primarily means that the source code
provided must be that code used to produce the binaries being distributed.
That source code must also be ``complete''. GPLv2~\S3's penultimate paragraph
explains in detail what is meant by ``complete''. In essence, it is all
the material that a programmer of average skill would need to actually use
the source code to produce the binaries she has received. Complete source
is required so that, if the licensee chooses, she should be able to
exercise her freedoms to modify and redistribute changes. Without the
complete source, it would not be possible to make changes that were
actually directly derived from the version received.
Furthermore, GPLv2~\S3 is defending against a tactic that has in fact been
seen in GPL enforcement. Under GPL, if you pay a high price for
a copy of GPL'd binaries (which comes with corresponding source, of
course), you have the freedom to redistribute that work at any fee you
choose, or not at all. Sometimes, companies attempt a GPL-violating
cozenage whereby they produce very specialized binaries (perhaps for
an obscure architecture). They then give source code that does
correspond, but withhold the ``incantations'' and build plans they
used to make that source compile into the specialized binaries.
Therefore, GPLv2~\S3 requires that the source code include ``meta-material'' like
scripts, interface definitions, and other material that is used to
``control compilation and installation'' of the binaries. In this
manner, those further down the distribution chain are assured that
they have the unabated freedom to build their own derivative works
from the sources provided.
Software distribution comes in many
forms. Embedded manufacturers, for example, have the freedom to put
GPL'd software into mobile devices with very tight memory and space
constraints. In such cases, putting the source right alongside the
binaries on the machine itself might not be an option. While it is
recommended that this be the default way that people comply with GPL, the
GPL does provide options when such distribution is infeasible.
\label{GPLv2s3-medium-customarily}
GPLv2~\S3, therefore, allows source code to be provided on any physical
``medium customarily used for software interchange.'' By design, this
phrase covers a broad spectrum --- the phrase seeks to pre-adapt to
changes in technology. When GPLv22 was first published in June
1991, distribution on magnetic tape was still common, and CD was
relatively new. By 2002, CD is the default. By 2007, DVD's were the
default. Now, it's common to give software on USB drives and SD card. This
language in the license must adapt with changing technology.
Meanwhile, the binding created by the word ``customarily'' is key. Many
incorrectly believe that distributing binary on CD and source on the
Internet is acceptable. In the corporate world in industrialized countries, it is indeed customary to
simply download a CDs' worth of data quickly. However, even today in the USA, many computer users are not connected to the Internet, and most people connected
to the Internet still have limited download speeds. Downloading
CDs full of data is not customary for them in the least. In some cities
in Africa, computers are becoming more common, but Internet connectivity
is still available only at a few centralized locations. Thus, the
``customs'' here are normalized for a worldwide userbase. Simply
providing source on the Internet --- while it is a kind, friendly and
useful thing to do --- is not usually sufficient.
Note, however, a major exception to this rule, given by the last paragraph
of GPLv2~\S3. \emph{If} distribution of the binary files is made only on the
Internet (i.e., ``from a designated place''), \emph{then} simply providing
the source code right alongside the binaries in the same place is
sufficient to comply with GPLv2~\S3.
\medskip
As is shown above, Under GPLv2~\S3(a), embedded manufacturers can put the
binaries on the device and ship the source code along on a CD\@. However,
sometimes this turns out to be too costly. Including a CD with every
device could prove too costly, and may practically (although not legally)
prohibit using GPL'd software. For this situation and others like it, GPLv2\S~3(b) is available.
GPLv2~\S3(b) allows a distributor of binaries to instead provide a written
offer for source code alongside those binaries. This is useful in two
specific ways. First, it may turn out that most users do not request the
source, and thus the cost of producing the CDs is saved --- a financial
and environmental windfall. In addition, along with a GPLv2~\S3(b) compliant
offer for source, a binary distributor might choose to \emph{also} give a
URL for source code. Many who would otherwise need a CD with source might
turn out to have those coveted high bandwidth connections, and are able to
download the source instead --- again yielding environmental and financial
windfalls.
However, note that regardless of how many users prefer to get the
source online, GPLv2~\S3(b) does place lasting long-term obligations on the
binary distributor. The binary distributor must be prepared to honor
that offer for source for three years and ship it out (just as they
would have had to do under GPLv2~\S3(a)) at a moment's notice when they
receive such a request. There is real organizational cost here:
support engineers must be trained how to route source requests, and
source CD images for every release version for the last three years
must be kept on hand to burn such CDs quickly. The requests might not
even come from actual customers; the offer for source must be valid
for ``any third party''.
That phrase is another place where some get confused --- thinking again
that full public distribution of source is required. The offer for source
must be valid for ``any third party'' because of the freedoms of
redistribution granted by GPLv2~\S\S1--2. A company may ship a binary image
and an offer for source to only one customer. However, under GPL, that
customer has the right to redistribute that software to the world if she
likes. When she does, that customer has an obligation to make sure that
those who receive the software from her can exercise their freedoms under
GPL --- including the freedom to modify, rebuild, and redistribute the
source code.
GPLv2~\S3(c) is created to save her some trouble, because by itself GPLv2~\S3(b)
would unfairly favor large companies. GPLv2~\S3(b) allows the
separation of the binary software from the key tool that people can use
to exercise their freedom. The GPL permits this separation because it is
good for redistributors, and those users who turn out not to need the
source. However, to ensure equal rights for all software users, anyone
along the distribution chain must have the right to get the source and
exercise those freedoms that require it.
Meanwhile, GPLv2~\S3(b)'s compromise primarily benefits companies who
distribute binary software commercially. Without GPLv2~\S3(c), that benefit
would be at the detriment of the companies' customers; the burden of
source code provision would be unfairly shifted to the companies'
customers. A customer, who had received binaries with a GPLv2~\S3(b)-compliant
offer, would be required under GPLv2 (sans GPLv2~\S3(c)) to acquire the source,
merely to give a copy of the software to a friend who needed it. GPLv2~\S3(c)
reshifts this burden to entity who benefits from GPLv2~\S3(b).
GPLv2~\S3(c) allows those who undertake \emph{noncommercial} distribution to
simply pass along a GPLv2~\S3(b)-compliant source code offer. The customer who
wishes to give a copy to her friend can now do so without provisioning the
source, as long as she gives that offer to her friend. By contrast, if
she wanted to go into business for herself selling CDs of that software,
she would have to acquire the source and either comply via GPLv2~\S3(a), or
write her own GPLv2~\S3(b)-compliant source offer.
This process is precisely the reason why a GPLv2~\S3(b) source offer must be
valid for all third parties. At the time the offer is made, there is no
way of knowing who might end up noncommercially receiving a copy of the
software. Companies who choose to comply via GPLv2~\S3(b) must thus be
prepared to honor all incoming source code requests. For this and the
many other additional necessary complications under GPLv2~\S\S3(b--c), it is
only rarely a better option than complying via GPLv2~\S3(a).
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\chapter{GPL's Implied Patent Grant}
\label{gpl-implied-patent-grant}
We digress again briefly from our section-by-section consideration of GPLv2
to consider the interaction between the terms of GPL and patent law. The
GPLv2, despite being silent with respect to patents, actually confers on its
licensees more rights to a licensor's patents than those licenses that
purport to address the issue. This is the case because patent law, under
the doctrine of implied license, gives to each distributee of a patented
article a license from the distributor to practice any patent claims owned
or held by the distributor that cover the distributed article. The
implied license also extends to any patent claims owned or held by the
distributor that cover ``reasonably contemplated uses'' of the patented
article. To quote the Federal Circuit Court of Appeals, the highest court
for patent cases other than the Supreme Court:
\begin{quotation}
Generally, when a seller sells a product without restriction, it in
effect promises the purchaser that in exchange for the price paid, it will
not interfere with the purchaser's full enjoyment of the product
purchased. The buyer has an implied license under any patents of the
seller that dominate the product or any uses of the product to which the
parties might reasonably contemplate the product will be put.
\end{quotation}
Hewlett-Packard Co. v. Repeat-O-Type Stencil Mfg. Corp., Inc., 123 F.3d
1445, 1451 (Fed. Cir. 1997).
Of course, Free Software is licensed, not sold, and there are indeed
restrictions placed on the licensee, but those differences are not likely
to prevent the application of the implied license doctrine to Free
Software, because software licensed under the GPL grants the licensee the
right to make, use, and sell the software, each of which are exclusive
rights of a patent holder. Therefore, although the GPLv2 does not expressly
grant the licensee the right to do those things under any patents the
licensor may have that cover the software or its reasonably contemplated
uses, by licensing the software under the GPLv2, the distributor impliedly
licenses those patents to the GPLv2 licensee with respect to the GPLv2'd
software.
An interesting issue regarding this implied patent license of GPLv2'd
software is what would be considered ``uses of the [software] to which
the parties might reasonably contemplate the product will be put.'' A
clever advocate may argue that the implied license granted by GPLv2 is
larger in scope than the express license in other Free Software
licenses with express patent grants, in that the patent license
clause of many of those other Free Software licenses are specifically
limited to the patent claims covered by the code as licensed by the patentee.
In contrast, a GPLv2 licensee, under the doctrine of implied patent license,
is free to practice any patent claims held by the licensor that cover
``reasonably contemplated uses'' of the GPL'd code, which may very well
include creation and distribution of derivative works since the GPL's terms,
under which the patented code is distributed, expressly permits such activity.
Further supporting this result is the Federal Circuit's pronouncement that
the recipient of a patented article has, not only an implied license to
make, use, and sell the article, but also an implied patent license to
repair the article to enable it to function properly, Bottom Line Mgmt.,
Inc. v. Pan Man, Inc., 228 F.3d 1352 (Fed. Cir. 2000). Additionally, the
Federal Circuit extended that rule to include any future recipients of the
patented article, not just the direct recipient from the distributor.
This theory comports well with the idea of Free Software, whereby software
is distributed amongst many entities within the community for the purpose
of constant evolution and improvement. In this way, the law of implied
patent license used by the GPLv2 ensures that the community mutually
benefits from the licensing of patents to any single community member.
Note that simply because GPLv2'd software has an implied patent license does
not mean that any patents held by a distributor of GPLv2'd code become
worthless. To the contrary, the patents are still valid and enforceable
against either:
\begin{enumerate}
\renewcommand{\theenumi}{\alph{enumi}}
\renewcommand{\labelenumi}{\textup{(\theenumi)}}
\item any software other than that licensed under the GPLv2 by the patent
holder, and
\item any party that does not comply with the GPLv2
with respect to the licensed software.
\end{enumerate}
\newcommand{\compB}{$\mathcal{B}$}
\newcommand{\compA}{$\mathcal{A}$}
For example, if Company \compA{} has a patent on advanced Web browsing, but
also licenses a Web browsing software program under the GPLv2, then it
cannot assert the patent against any party based on that party's use of
Company \compA{}'s GPL'ed Web browsing software program, or on that party's
creation and use of derivative works of that GPL'ed program. However, if a
party uses that program without
complying with the GPLv2, then Company \compA{} can assert both copyright
infringement claims against the non-GPLv2-compliant party and
infringement of the patent, because the implied patent license only
extends to use of the software in accordance with the GPLv2. Further, if
Company \compB{} distributes a competitive advanced Web browsing program
that is not a derivative work of Company \compA{}'s GPL'ed Web browsing software
program, Company \compA{} is free to assert its patent against any user or
distributor of that product. It is irrelevant whether Company \compB's
program is also distributed under the GPLv2, as Company \compB{} can not grant
implied licenses to Company \compA's patent.
This result also reassures companies that they need not fear losing their
proprietary value in patents to competitors through the GPLv2 implied patent
license, as only those competitors who adopt and comply with the GPLv2's
terms can benefit from the implied patent license. To continue the
example above, Company \compB{} does not receive a free ride on Company
\compA's patent, as Company \compB{} has not licensed-in and then
redistributed Company A's advanced Web browser under the GPLv2. If Company
\compB{} does do that, however, Company \compA{} still has not lost
competitive advantage against Company \compB{}, as Company \compB{} must then,
when it re-distributes Company \compA's program, grant an implied license
to any of its patents that cover the program. Further, if Company \compB{}
relicenses an improved version of Company A's program, it must do so under
the GPLv2, meaning that any patents it holds that cover the improved version
are impliedly licensed to any licensee. As such, the only way Company
\compB{} can benefit from Company \compA's implied patent license, is if it,
itself, distributes Company \compA's software program and grants an
implied patent license to any of its patents that cover that program.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\chapter{Defending Freedom on Many Fronts}
Chapters~\ref{run-and-verbatim} and~\ref{source-and-binary} presented the
core freedom-defending provisions of GPLv2\@, which are in GPLv2~\S\S0--3.
GPLv2\S\S~4--7 of the GPLv2 are designed to ensure that GPLv2~\S\S0--3 are
not infringed, are enforceable, are kept to the confines of copyright law but
also not trumped by other copyright agreements or components of other
entirely separate legal systems. In short, while GPLv2~\S\S0--3 are the parts
of the license that defend the freedoms of users and programmers,
GPLv2~\S\S4--7 are the parts of the license that keep the playing field clear
so that \S\S~0--3 can do their jobs.
\section{GPLv2~\S4: Termination on Violation}
\label{GPLv2s4}
GPLv2~\S4 is GPLv2's termination clause. Upon first examination, it seems
strange that a license with the goal of defending users' and programmers'
freedoms for perpetuity in an irrevocable way would have such a clause.
However, upon further examination, the difference between irrevocability
and this termination clause becomes clear.
The GPL is irrevocable in the sense that once a copyright holder grants
rights for someone to copy, modify and redistribute the software under terms
of the GPL, they cannot later revoke that grant. Since the GPL has no
provision allowing the copyright holder to take such a prerogative, the
license is granted as long as the copyright remains in effect.\footnote{In
the USA, due to unfortunate legislation, the length of copyright is nearly
perpetual, even though the Constitution forbids perpetual copyright.} The
copyright holders have the right to relicense the same work under different
licenses (see Section~\ref{Proprietary Relicensing} of this tutorial), or to
stop distributing the GPLv2'd version (assuming GPLv2~\S3(b) was never used),
but they may not revoke the rights under GPLv2 already granted.
In fact, when an entity looses their right to copy, modify and distribute
GPL'd software, it is because of their \emph{own actions}, not that of the
copyright holder. The copyright holder does not decided when GPLv2~\S4
termination occurs (if ever); rather, the actions of the licensee determine
that.
Under copyright law, the GPL has granted various rights and freedoms to
the licensee to perform specific types of copying, modification, and
redistribution. By default, all other types of copying, modification, and
redistribution are prohibited. GPLv2~\S4 says that if you undertake any of
those other types (e.g., redistributing binary-only in violation of GPLv2~\S3),
then all rights under the license --- even those otherwise permitted for
those who have not violated --- terminate automatically.
GPLv2~\S4 makes GPLv2 enforceable. If licensees fail to adhere to the
license, then they are stuck without any permission under to engage in
activities covered by copyright law. They must completely cease and desist
from all copying, modification and distribution of the GPL'd software.
At that point, violating licensees must gain the forgiveness of the copyright
holders to have their rights restored. Alternatively, the violators could
negotiate another agreement, separate from GPL, with the copyright
holder. Both are common practice, although
\tutorialpartsplit{as discussed in \textit{A Practical Guide to GPL
Compliance}, there are }{Chapter~\ref{compliance-understanding-whos-enforcing}
explains further } key differences between these two very different uses of GPL.
\section{GPLv2~\S5: Acceptance, Copyright Style}
\label{GPLv2s5}
GPLv2~\S5 brings us to perhaps the most fundamental misconception and common
confusion about GPLv2\@. Because of the prevalence of proprietary software,
most users, programmers, and lawyers alike tend to be more familiar with
EULAs. EULAs are believed by their authors to be contracts, requiring
formal agreement between the licensee and the software distributor to be
valid. This has led to mechanisms like ``shrink-wrap'' and ``click-wrap''
as mechanisms to perform acceptance ceremonies with EULAs.
The GPL does not need contract law to ``transfer rights.'' Usually, no rights
are transfered between parties. By contrast, the GPL is primarily a permission
slip to undertake activities that would otherwise have been prohibited
by copyright law. As such, GPL needs no acceptance ceremony; the
licensee is not even required to accept the license.
However, without the GPL, the activities of copying, modifying and
distributing the software would have otherwise been prohibited. So, the
GPL says that you only accepted the license by undertaking activities that
you would have otherwise been prohibited without your license under GPL\@.
This is a certainly subtle point, and requires a mindset quite different
from the contractual approach taken by EULA authors.
An interesting side benefit to GPLv2~\S5 is that the bulk of users of Free
Software are not required to accept the license. Undertaking fair and
unregulated use of the work, for example, does not bind you to the GPL,
since you are not engaging in activity that is otherwise controlled by
copyright law. Only when you engage in those activities that might have an
impact on the freedom of others does license acceptance occur, and the
terms begin to bind you to fair and equitable sharing of the software. In
other words, the GPL only kicks in when it needs to for the sake of
freedom.
While GPL is by default a copyright license, it is certainly still possible
to consider GPL as a contract as well. For example, some distributors chose
to ``wrap'' their software in an acceptance ceremony to GPL, and nothing in
GPL prohibits that use. Furthermore, the ruling in \textit{Jacobsen
v. Katzer, 535 F.3d 1373, 1380 (Fed.Cir.2008)} indicates that \textbf{both}
copyright and contractual remedies may be sought by a copyright holder
seeking to enforce a license designed to uphold software freedom.
\section{Using GPL Both as a Contract and Copyright License}
\section{GPLv2~\S6: GPL, My One and Only}
\label{GPLv2s6}
A point that was glossed over in Section~\ref{GPLv2s4}'s discussion of GPLv2~\S4
was the irrevocable nature of the GPL\@. The GPLv2 is indeed irrevocable,
and it is made so formally by GPLv2~\S6.
The first sentence in GPLv2~\S6 ensures that as software propagates down the
distribution chain, that each licensor can pass along the license to each
new licensee. Under GPLv2~\S6, the act of distributing automatically grants a
license from the original licensor to the next recipient. This creates a
chain of grants that ensure that everyone in the distribution has rights
under the GPLv2\@. In a mathematical sense, this bounds the bottom ---
making sure that future licensees get no fewer rights than the licensee before.
The second sentence of GPLv2~\S6 does the opposite; it bounds from the top. It
prohibits any licensor along the distribution chain from placing
additional restrictions on the user. In other words, no additional
requirements may trump the rights and freedoms given by GPLv2\@.
The final sentence of GPLv2~\S6 makes it abundantly clear that no individual
entity in the distribution chain is responsible for the compliance of any
other. This is particularly important for noncommercial users who have
passed along a source offer under GPLv2~\S3(c), as they cannot be assured that
the issuer of the offer will honor their GPLv2~\S3 obligations.
In short, GPLv2~\S6 says that your license for the software is your one and
only copyright license allowing you to copy, modify and distribute the
software.
\section{GPLv2~\S7: ``Give Software Liberty or Give It Death!''}
\label{GPLv2s7}
In essence, GPLv2~\S7 is a verbosely worded way of saying for non-copyright
systems what GPLv2~\S6 says for copyright. If there exists any reason that a
distributor knows of that would prohibit later licensees from exercising
their full rights under GPL, then distribution is prohibited.
Originally, this was designed as the title of this section suggests --- as
a last ditch effort to make sure that freedom was upheld. However, in
modern times, it has come to give much more. Now that the body of GPL'd
software is so large, patent holders who would want to be distributors of
GPL'd software have a tough choice. They must choose between avoiding
distribution of GPL'd software that exercises the teachings of their
patents, or grant a royalty-free, irrevocable, non-exclusive license to
those patents. Many companies have chosen the latter.
Thus, GPLv2~\S7 rarely gives software death by stopping its distribution.
Instead, it is inspiring patent holders to share their patents in the same
freedom-defending way that they share their copyrighted works.
\section{GPLv2~\S8: Excluding Problematic Jurisdictions}
\label{GPLv2s8}
GPLv2~\S8 is rarely used by copyright holders. Its intention is that if a
particular country, say Unfreedonia, grants particular patents or allows
copyrighted interfaces (no country to our knowledge even permits those
yet), that the GPLv2'd software can continue in free and unabated
distribution in the countries where such controls do not exist.
As far as is currently known, GPLv2~\S8 has very rarely been formally used by
copyright holders. Admittedly, some have used GPLv2~\S8 to explain various
odd special topics of distribution (usually related in some way to
GPLv2~\S7). However, generally speaking, this section is not proven
particularly useful in the more than two decades of GPLv2 history.
Meanwhile, despite many calls by the FSF (and others) for those licensors who
explicitly use this section to come forward and explain their reasoning, no
one ever did. Furthermore, research conducted during the GPLv3 drafting
process found exactly one licensor who had invoked this section to add an
explicit geographical distribution limitation, and the reasoning for that one
invocation was not fitting with FSF's intended spirit of GPLv2~\S8. As such,
GPLv2~\S8 was not included at all in GPLv3.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\chapter{Odds, Ends, and Absolutely No Warranty}
GPLv2~\S\S0--7 constitute the freedom-defending terms of the GPLv2. The remainder
of the GPLv2 handles administrivia and issues concerning warranties and
liability.
\section{GPLv2~\S9: FSF as Stewards of GPL}
\label{GPLv2s9}
FSF reserves the exclusive right to publish future versions of the GPL\@;
GPLv2~\S9 expresses this. While the stewardship of the copyrights on the body
of GPL'd software around the world is shared among thousands of
individuals and organizations, the license itself needs a single steward.
Forking of the code is often regrettable but basically innocuous. Forking
of licensing is disastrous.
(Chapter~\ref{tale-of-two-copylefts} discusses more about the various
versions of GPL.)
\section{GPLv2~\S10: Relicensing Permitted}
\label{GPLv2s10}
GPLv2~\S10 reminds the licensee of what is already implied by the nature of
copyright law. Namely, the copyright holder of a particular software
program has the prerogative to grant alternative agreements under separate
copyright licenses.
\section{GPLv2~\S11: No Warranty}
\label{GPLv2s11}
Most warranty disclaimer language shout at you. The
\href{http://www.law.cornell.edu/ucc/2/2-316}{Uniform Commercial
Code~\S2-316} requires that disclaimers of warranty be ``conspicuous''.
There is apparently general acceptance that \textsc{all caps} is the
preferred way to make something conspicuous, and that has over decades worked
its way into the voodoo tradition of warranty disclaimer writing.
That said, there is admittedly some authority under USA law suggesting that
effective warranty disclaimers that conspicuousness can be established by
capitalization and is absent when a disclaimer has the same typeface as the
terms surrounding it (see \textit{Stevenson v.~TRW, Inc.}, 987 F.2d 288, 296
(5th Cir.~1993)). While GPLv3's drafters doubted that such authority would
apply to copyright licenses like the GPL, the FSF has nevertheless left
warranty and related disclaimers in \textsc{all caps} throughout all versions
of GPL\@\footnote{One of the authors of this tutorial, Bradley M.~Kuhn, has
often suggested the aesthetically preferable compromise of a
\textsc{specifically designed ``small caps'' font, such as this one, as an
alternative to} WRITING IN ALL CAPS IN THE DEFAULT FONT (LIKE THIS),
since the latter adds more ugliness than conspicuousness. Kuhn once
engaged in reversion war with a lawyer who disagreed, but that lawyer never
answered Kuhn's requests for case law that argues THIS IS INHERENTLY MORE
CONSPICUOUS \textsc{Than this is}.}.
Some have argued the GPL is unenforceable in some jurisdictions because
its disclaimer of warranties is impermissibly broad. However, GPLv2~\S11
contains a jurisdictional savings provision, which states that it is to be
interpreted only as broadly as allowed by applicable law. Such a
provision ensures that both it, and the entire GPL, is enforceable in any
jurisdiction, regardless of any particular law regarding the
permissibility of certain warranty disclaimers.
Finally, one important point to remember when reading GPLv2~\S11 is that GPLv2~\S1
permits the sale of warranty as an additional service, which GPLv2~\S11 affirms.
\section{GPLv2~\S12: Limitation of Liability}
\label{GPLv2s12}
There are many types of warranties, and in some jurisdictions some of them
cannot be disclaimed. Therefore, usually agreements will have both a
warranty disclaimer and a limitation of liability, as we have in GPLv2~\S12.
GPLv2~\S11 thus gets rid of all implied warranties that can legally be
disavowed. GPLv2~\S12, in turn, limits the liability of the actor for any
warranties that cannot legally be disclaimed in a particular jurisdiction.
Again, some have argued the GPL is unenforceable in some jurisdictions
because its limitation of liability is impermissibly broad. However, \S
12, just like its sister, GPLv2~\S11, contains a jurisdictional savings
provision, which states that it is to be interpreted only as broadly as
allowed by applicable law. As stated above, such a provision ensures that
both GPLv2~\S12, and the entire GPL, is enforceable in any jurisdiction,
regardless of any particular law regarding the permissibility of limiting
liability.
So end the terms and conditions of the GNU General Public License.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\chapter{GPL Version 3}
\label{GPLv3}
This chapter discussed the text of GPLv3. Much of this material herein
includes text that was adapted (with permission) from text that FSF
originally published as part of the so-called ``rationale documents'' for the
various discussion drafts of GPLv3.
The FSF ran a somewhat public process to develop GPLv3, and it was the first
attempt of its kind to develop a Free Software license this way. Ultimately,
RMS was the primary author of GPLv3, but he listened to feedback from all
sorts of individuals and even for-profit companies. Nevertheless, in
attempting to understand GPLv3 after the fact, the materials available from
the GPLv3 process have a somewhat ``drinking from the firehose'' effect.
This chapter seeks to explain GPLv3 to newcomers, who perhaps are familiar
with GPLv2 and who did not participate in the GPLv3 process.
Those who wish to drink from the firehose and take a diachronic approach to
GPLv3 study by reading the step-by-step public drafting process GPLv3 (which
occurred from Monday 16 January 2006 through Monday 19 November 2007) should
visit \url{http://gplv3.fsf.org/}.
\section{Understanding GPLv3 As An Upgraded GPLv2}
Ultimately, GPLv2 and GPLv3 co-exist as active licenses in regular use. As
discussed in Chapter\~ref{tale-of-two-copylefts}, GPLv1 was never regularly
used alongside GPLv2. However, given GPLv2's widespread popularity and
existing longevity by the time GPLv3 was published, it is not surprising that
some licensors still prefer GPLv2-only or GPLv2-or-later. GPLv3 gained major
adoption by many projects, old and new, but many projects have not upgraded
due to (in some cases) mere laziness and (in other cases) policy preference
for some of GPLv2's terms and/or policy opposition to GPLv3's terms.
Given this ``two GPLs world'' is reality, it makes sense to consider GPLv3 in
terms of how it differs from GPLv2. Also, most of the best GPL experts in
the world must deal regularly with both licenses, and admittedly have decades
of experience of GPLv2 while the most experience with GPLv3 that's possible
is by default less than a decade. These two factors usually cause even new
students of GPL to start with GPLv2 and move on to GPLv3, and this tutorial
follows that pattern.
Overall, the changes made in GPLv3 admittedly \textit{increased} the
complexity of the license. The FSF stated at the start of the GPLv3 process
that they would have liked to oblige those who have asked for a simpler and
shorter GPL\@. Ultimately, the FSF gave priority to making GPLv3 a better
copyleft in the spirit of past GPL's. Obsession for concision should never
trump software freedom.
The FSF had many different, important goals in seeking to upgrade to GPLv3.
However, one important goal that is often lost in the discussion of policy
minutia is a rather simple but important issue. Namely, FSF sought to assure
that GPLv3 was more easily internationalized than GPLv2. In particular, the
FSF sought to ease interpretation of GPL in other countries by replacement of
USA-centric\footnote{See Section~\ref{non-usa-copyright} of this tutorial for
a brief discussion about non-USA copyright systems.} copyright phrases and
wording with neutral terminology rooted in description of behavior rather
than specific statue. As can be seen in the section-by-section discussion of
GPLv3 that follows, nearly every section had changes related to issues of
internationalization.
\section{GPLv3~\S0: Giving In On ``Defined Terms''}
One of lawyers' most common complaints about GPLv2 is that defined terms in
the document appear throughout. Most licenses define terms up-front.
However, GPL was always designed both as a document that should be easily
understood both by lawyers and by software developers: it is a document
designed to give freedom to software developers and users, and therefore it
should be comprehensible to that constituency.
Interestingly enough, one coauthor of this tutorial who is both a lawyer and
a developer pointed out that in law school, she understood defined terms more
quickly than other law students precisely because of her programming
background. For developers, having \verb0#define0 (in the C programming
language) or other types of constants and/or macros that automatically expand
in the place where they are used is second nature. As such, adding a defined
terms section was not terribly problematic for developers, and thus GPLv3
adds one. Most of these defined terms are somewhat straightforward and bring
forward better worded definitions from GPLv2. Herein, this tutorial
discusses a few of the new ones.
GPLv3~\S0 includes definitions of five new terms not found in any form in
GPLv2: ``modify'' ``covered work'', ``propagate'', ``convey'', and
``Appropriate Legal Notices''.
\subsection{Modify and the Work Based on the Program}
GPLv2 included a defined term, ``work based on the Program'', but also used
the term ``modify'' and ``based on'' throughout the license. GPLv2's ``work
based on the Program'' definition made use of a legal term of art,
``derivative work'', which is peculiar to USA copyright law. However,
ironically, the most criticism of USA-specific legal terminology in GPLv2's
``work based on the Program'' definition historically came not primarily from
readers outside the USA, but from those within it\footnote{The FSF noted in
that it did not generally agree with these views, and expressed puzzlement
by the energy with which they were expressed, given the existence of many
other, more difficult legal issues implicated by the GPL. Nevertheless,
the FSF argued that it made sense to eliminate usage of local copyright
terminology to good effect.}. Admittedly, even though differently-labeled
concepts corresponding to the derivative work are recognized in all copyright
law systems, these counterpart concepts might differ to some degree in scope
and breadth from the USA derivative work.
The goal and intention of GPLv2 was always to cover all rights governed by
relevant copyright law, in the USA and elsewhere. GPLv3 therefore takes the
task of internationalizing the license further by removing references to
derivative works and by providing a more globally useful definition. The new
definition returns to the common elements of copyright law. Copyright
holders of works of software have the exclusive right to form new works by
modification of the original --- a right that may be expressed in various
ways in different legal systems. GPLv3 operates to grant this right to
successive generations of users (particularly through the copyleft conditions
set forth in GPLv3~\S5, as described later in this tutorial in its
\S~\ref{GPLv3s5}). Here in GPLv3~\S0, ``modify'' refers to basic copyright
rights, and then this definition of ``modify'' is used to define ``modified
version of'' and ``work based on,'' as synonyms.
\section{The Covered Work}
GPLv3 uses a common license drafting technique of building upon simpler
definitions to make complex ones. The Program is a defined term found
throughout GPLv2, and the word ``covered'' and the phrase ``covered by this
license'' are used in tandem with the Program in GPLv2, but not as part of a
definition. GPLv3 offers a single term ``covered work'', which enables some
of the wording in GPLv3 to be simpler and clearer than its GPLv2
counterparts.
\section{Propagate}
The term ``propagate'' serves two purposes. First, ``propagate'' provides a
simple and convenient means for distinguishing between the kinds of uses of a
work that GPL imposes conditions on and the kinds of uses that GPL does not
(for the most part) impose conditions on.
Second, ``propagate'' helps globalize GPL in its wording and effect. When a
work is GPL'd, the copyright law of some particular country will govern
certain legal issues arising under the license. A term like ``distribute''
(or its equivalent in languages other than English) is used in several
national copyright statutes. Yet, practical experience with GPLv2 revealed
the awkwardness of using the term ``distribution'' in a license intended for
global use: the scope of ``distribution'' in the copyright context can differ
from country to country. The GPL never necessarily intended the specific
meaning of ``distribution'' that exists under USA (or any other country's)
copyright law.
Indeed, even within a single country and language, the term distribution may
be ambiguous; as a legal term of art, distribution varies significantly in
meaning among those countries that recognize it. For example, comments
during GPLv3's drafting process indicated that in at least one country,
distribution may not include network transfers of software but may include
interdepartmental transfers of physical copies within an organization.
Meanwhile, the copyright laws of many countries, as well as certain
international copyright treaties, recognize ``making available to the
public'' or ``communication to the public'' as one of the exclusive rights of
copyright holders.
Therefore, the GPL defines the term ``propagate'' by reference to activities
that require permission under ``applicable copyright law'', but excludes
execution and private modification from the definition. GPLv3's definition
also gives examples of activities that may be included within ``propagation''
but it also makes clear that, under the copyright laws of a given country,
``propagation'' may include other activities as well.
Thus, propagation is defined by behavior, and not by categories drawn from
some particular national copyright statute. This helps not only with
internationalization, but also factually-based terminology aids in
developers' and users' understanding of GPL\@.
\section{Convey}
Further to this point, a subset of propagate --- ``convey'' --- is defined.
Conveying includes activities that constitute propagation of copies to
others. As with the definition of propagate, GPLv3 thus addresses transfers
of copies of software in behavioral rather than statutory terms.
\section{Appropriate Legal Notices}
GPLv2 used the term ``appropriate copyright notice and disclaimer of
warranty'' in two places, which is a rather bulk term. Also, experience with
GPLv2 and other licenses that grant software freedom showed throughout the
1990s that the scope of types of notices that need preservation upon
conveyance were more broad that merely the copyright notices. The
Appropriate Legal Notice definition consolidates the material that GPLv2
traditionally required preserved into one definition.
\section{Other Defined Terms}
Note finally that not all defined terms in GPLv3 appear in GPLv3~\S0.
Specifically, those defined terms that are confined in use to a single
section are defined in the section in which they are used, and GPLv3~\S1
contains those definitions focused on source code. In this tutorial, those
defined terms are discussed in the section where they are defined and/or
used.
\section{GPLv3~\S1: Understanding CCS}
Ensuring that users have the source code to the software they receive and the
freedom to modify remains the paramount right embodied in the Free Software
Definition (found in \S~\ref{Free Software Definition} of this tutorial). As
such, GPLv3~\S1 is likely one of the most important sections of GPLv3, as it
contains all the defined terms related to this important software freedom.
\subsection{Source Code Definition}
First, GPLv3\~S1 retains GPLv2's definition of ``source code'' and adds an
explicit definition of ``object code'' as ``any non-source version of a
work''. Object code is not restricted to a narrow technical meaning and is
understood broadly to include any form of the work other than the preferred
form for making modifications to it. Object code therefore includes any kind
of transformed version of source code, such as bytecode or minified
Javascript. The definition of object code also ensures that licensees cannot
escape their obligations under the GPL by resorting to shrouded source or
obfuscated programming.
\subsection{CCS Definition}
The definition of CCS\footnote{Note that the preferred term for those who
work regularly with both GPLv2 and GPLv3 is ``Complete Corresponding
Source'', abbreviated to ``CCS''. Admittedly, the word ``complete'' no
longer appears in GPLv3 (which uses the word ``all'' instead). However,
both GPLv2 and the early drafts of GPLv3 itself used the word ``complete'',
and early GPLv3 drafts even called this defined term ``Complete
Corresponding Source''. Meanwhile, use of the acronym ``CCS'' (sometimes,
``C\&CS'') was so widespread among GPL enforcers that its use continues
even though GPLv3-focused experts tend to say just the defined term of
``Corresponding Source''.}, or, as GPLv3 officially calls it,
``Corresponding Source'' in GPLv3~\S1\P4 is possibly the most complex
definition in the license.
The CCS definition is broad so as to protect users' exercise of their rights
under the GPL\@. The definition includes with particular examples to remove
any doubt that they are to be considered CCS\@. GPLv3 seeks to make it
completely clear that a licensee cannot avoid complying with the requirements
of the GPL by dynamically linking a subprogram component to the original
version of a program. The example also clarifies that the shared libraries
and dynamically linked subprograms that are included in Corresponding Source
are those that the work is ``specifically'' designed to require, which
clarifies that they do not include libraries invoked by the work that can be
readily substituted by other existing implementations. While copyleft
advocates never doubted this was required under GPLv2's definition of CCS,
making it abundantly clear with an extra example.
GPL, as always, seeks to ensure users are truly in a position to install and
run their modified versions of the program; the CCS definition is designed to
be expansive to ensure this software freedom. However, although the
definition of CCS is expansive, it is not sufficient to protect users'
freedoms in many circumstances. For example, a GPL'd program, or a modified
version of such a program, might be locked-down and restricted. The
requirements in GPLv3~\S6 (discussed in Section~\ref{GPLv3s6} of this
tutorial) handle that issue. (Early drafts of GPLv3 included those
requirements in the definition of CCS; however, given that the lock-down
issue only comes up in distribution of object code, it is more logical to
place those requirements with the parts of GPLv3 dealing directly with object
code distribution).
The penultimate paragraph in GPLv3\S2 notes that GPLv3's CCS definition does
not require source that can be automatically generated. Many code
generators, preprocessors and take source code as input and sometimes even
have output that is still source code. Source code should always be whatever
the original programmer preferred to modify.
GPLv3\S1's final paragraph removes any ambiguity about what should be done on
source-only distributions. Specifically, the right to convey source code
that does not compile, does not work, or otherwise is experimental
in-progress work is fully permitted, \textit{provided that} no object code
form is conveyed as well. Indeed, when combined with the permissions in
GPLv3\S~5, it is clear that if one conveys \textit{only} source code, one can
never be required to provide more than that. One always has the right to
modify a source code work by deleting any part of it, and there can be no
requirement that free software source code be a whole functioning program.
\subsection{The System Library Exception}
The previous section skipped over one part of the CCS definition, the
so-called system library exception. The ``System Libraries'' definition (and
the ``Standard Interface'' and ``Major Component'' definitions, which it
includes) are designed as part
to permit certain distribution arrangements that are considered reasonable by
copyleft advocates. The system library exception is designed to allow
copylefted software to link with these libraries when such linking would hurt
software freedom more than it would hurt proprietary software.
The system library exception has two parts. Part (a) rewords the GPLv2
exception for clarity replaces GPLv2's words ``unless that component itself
accompanies the executable'' with ``which is not part of the Major
Component''. The goal here is to not require disclosure of source code of
certain libraries, such as necessary Microsoft Windows DLLs (which aren't
part of Windows' kernel but accompany it) that are required for functioning
of copylefted programs compiled for Windows.
However, in isolation, (a) would be too permissive, as it would sometimes
allowing distributors to evade important GPL requirements. Part (b) reigns
in (a). Specifically, (b) specifies only a few functionalities that a the
system library may provide and still qualify for the exception. The goal is
to ensure system libraries are truly adjunct to a major essential operating
system component, compiler, or interpreter. The more low-level the
functionality provided by the library, the more likely it is to be qualified
for this exception.
Admittedly, the system library exception is a frequently discussed topic of
obsessed GPL theorists. The amount that has been written on the system
library exception (both the GPLv2 and GPLv3 versions of it), if included
herein, could easily increase this section of the tutorial to a length
greater than all the others.
Like any exception to the copyleft requirements of GPL, would-be GPL
violators frequently look to the system library exception as a potential
software freedom circumvention technique. When considering whether or not a
library qualifies for the system library exception, here is a pragmatic
thesis to consider, based on the combined decades of experience in GPL
interpretation of this tutorial's authors: the harder and more strained the
reader must study and read the system library exception, the more likely it
is that the library in question does not qualify for it.
\section{GPLv3~\S2: Basic Permissions}
GPLv3~\S2 can roughly be considered as an equivalent to GPLv2~\S0 (discussed
in \S~\ref{GPLv2s0} of this tutorial). However, the usual style of
improvements found in GPLv3 are found here as well. For example, the first
sentence of GPLv3~\S2 furthers the goal internationalization. Under the
copyright laws of some countries, it may be necessary for a copyright license
to include an explicit provision setting forth the duration of the rights
being granted. In other countries, including the USA, such a provision is
unnecessary but permissible.
GPLv3~\S2\P1 also acknowledges that licensees under the GPL enjoy rights of
copyright fair use, or the equivalent under applicable law. These rights are
compatible with, and not in conflict with, the freedoms that the GPL seeks to
protect, and the GPL cannot and should not restrict them.
However, note that (sadly to some copyleft advocates) the unlimited freedom
to run is confined to the \textit{unmodified} Program. This confinement is
unfortunately necessary since Programs that do not qualify as a User Product
in GPLv3~\S6 (see \S~\ref{user-product} in this tutorial) might have certain
unfortunate restrictions on the freedom to run\footnote{See
\S~ref{freedom-to-run} of this tutorial for the details on ``the freedom to
run''.}
GPLv3~\S2\P2 distinguishes between activities of a licensee that are
permitted without limitation and activities that trigger additional
requirements. Specifically, GPLv3~\S2\P2 guarantees the basic freedoms of
privately modifying and running the program.
Also, GPLv3~\S2\P2 gives an explicit permission for a client to provide a
copy of its modified software to a contractor exclusively for that contractor
to modify it further, or run it, on behalf of the client. However, the
client can \textit{only} exercise this control over its own copyrighted
changes to the GPL-covered program. The parts of the program it obtained
from other contributors must be provided to the contractor with the usual GPL
freedoms. Thus, GPLv3 permits users to convey covered works to contractors
operating exclusively on the users' behalf, under the users' direction and
control, and to require the contractors to keep the users' copyrighted
changes confidential, but \textit{only if} the contractor is limited to acting
on the users' behalf (just as the users' employees would have to act).
The strict conditions in this ``contractors provision'' are needed so that it
cannot be twisted to fit other activities, such as making a program available
to downstream users or customers. By making the limits on this provision
very narrow, GPLv3 ensures that, in all other cases, contractors gets the
full freedoms of the GPL that they deserve.
GPLv3~\S2's final paragraph includes an explicit prohibition of sublicensing.
This provision ensures that GPL enforcement is always by the copyright
holder. Usually, sublicensing is regarded as a practical convenience or
necessity for the licensee, to avoid having to negotiate a license with each
licensor in a chain of distribution. The GPL solves this problem in another
way --- through its automatic licensing provision found in GPLv3\~S10 (which
is discussed in more detail in \S\~ref{GPLv3s10} of this tutorial).
\section{GPLv3's views on DRM and Device Lock-Down}
\label{GPLv3-drm}
The issues of DRM, device lock-down and encryption key disclosure were the
most hotly debated during the GPLv3 process. FSF's views on this were sadly
frequently misunderstood and, comparing the provisions related to these
issues in the earliest drafts of GPLv3 to the final version of GPLv3 shows
the FSF's willingness to compromise on tactical issues to reach the larger
goal of software freedom.
Specifically, GPLv3 introduced provisions that respond to the growing
practice of distributing GPL-covered programs in devices that employ
technical means to restrict users from installing and running modified
versions. This practice thwarts the expectations of developers and users
alike, because the right to modify is one of the core freedoms the GPL is
designed to secure.
Technological measures to defeat users' rights. These measures are often
described by such Orwellian phrases, such as ``digital rights management,''
which actually means limitation or outright destruction of users' legal
rights, or ``trusted computing,'' which actually means selling people
computers they cannot trust. However, these measures are alike in one basic
respect. They all employ technical means to turn the system of copyright law
(where the powers of the copyright holder are limited exceptions to general
freedom) into a virtual prison, where everything not specifically permitted
is utterly forbidden. This system of ``para-copyright'' was created well
after GPLv2 was written --- initially through legislation in the USA and the
EU, and later in other jurisdictions as well. This legislation creates
serious civil or even criminal penalties to escape from these restrictions
(commonly and aptly called ``jail-breaking a device''), even where the
purpose in doing so is to restore the users' legal rights that the technology
wrongfully prevents them from exercising.
GPLv2 did not address the use of technical measures to take back the rights
that the GPL granted, because such measures did not exist in 1991, and would
have been irrelevant to the forms in which software was then delivered to
users. GPLv3 addresses these issues, particularly because copylefted
software is ever more widely embedded in devices that impose technical
limitations on the user's freedom to change it.
However, FSF always made a clear distinction to avoid conflating these
``lock-down'' measures with legitimate applications that give users control,
as by enabling them to choose higher levels of system or data security within
their networks, or by allowing them to protect the security of their
communications using keys they can generate or copy to other devices for
sending or receiving messages. Such technologies present no obstacles to
software freedom and the goals of copyleft.
The public GPLv3 drafting process sought to balance these positions of
copyleft advocates with various desperate views of the larger
Free-Software-using community. Ultimately, FSF compromised to the GPLv3\S3
and GPLv3\S6 provisions that, taken together, are a minimalist set of terms
sufficient to protect the software freedom against the threat of invasive
para-copyright.
The compromises made were ultimately quite reasonable. The primary one is
embodied in GPLv3\S6's ``User Product'' definition (see \S~\ref{user-product}
in this tutorial for details). Additionally, some readers of early GPLv3
drafts seem to have assumed GPLv3 contained a blanket prohibition on DRM; but
it does not. In fact, no part of GPLv3 forbids DRM regarding non-GPL'd
works; rather, GPLv3 forbids the use of DRM specifically to lock-down
restrictions on users' ability to install modified versions of the GPL'd
software itself, but again, \textit{only} with regard to User Products.
Large enterprise users of free software often contract with non-employee
developers, often working offsite, to make modifications intended for
the user's private or internal use, and often arrange with other
companies to operate their data centers. Whether GPLv2 permits these
activities is not clear and may depend on variations in copyright law.
The practices seem basically harmless, so we have decided to make it
clear they are permitted.
\section{GPLv3~\S3: What Hath DMCA Wrought}
\label{GPLv3s3}
As discussed in \S~\ref{software-and-non-copyright} of this tutorial,
\href{http://www.law.cornell.edu/uscode/text/17/1201}{17 USC~\S1201} and
relate sections\footnote{These sections of the USC are often referred to as
the ``Digital Millennium Copyright Act'', or ``DMCA'', as that was the name
of the bill that so-modified these sections of the USC\@.} prohibits users
from circumventing technological measures that implement DRM\@. Since this
is part of copyright law and the GPL is primarily a copyright license, and
since what the DMCA calls ``circumvention'' is simply ``modifying the
software'' under the GPL, GPLv3 must disclaim such anti-circumvention
provisions are not applicable to the GPLv3'd software. GPLv3\S3 shields
users from being subjected to liability under anti-circumvention law for
exercising their rights under the GPL, so far as the GPL can do so.
First, GPLv3\S3\P1 declares that no GPL'd program is part of an effective
technological protection measure, regardless of what the program does. Early
drafts of GPLv3\S3\P1 referred directly to the DMCA, but the final version
instead includes instead an international legal reference to
anticircumvention laws enacted pursuant to the 1996 WIPO treaty and any
similar laws. Lawyers outside the USA worried that a USA statutory reference
could be read as indicating a choice for application of USA law to the
license as a whole. While the FSF did not necessarily agree with that view,
the FSF decided anyway to refer to the WIPO treaty rather than DMCA, since
several national anticircumvention laws were (or will likely be) structured
more similarly to the anticircumvention provisions of the DMCA in their
implementation of WIPO\@. Furthermore, the addition of ``or similar laws''
provides an appropriate catch-all.
Furthermore, GPLv3\S3\P2 states precisely that a conveying party waives the
power to forbid circumvention of technological measures only to the extent
that such circumvention is accomplished through the exercise of GPL rights in
the conveyed work. GPLv3\S3\P2 makes clear that the referenced ``legal
rights'' are specifically rights arising under anticircumvention law. and
refers to both the conveying party's rights and to third party rights, as in
some cases the conveying party will also be the party legally empowered to
enforce or invoke rights arising under anticircumvention law.
These disclaimers by each licensor of any intention to use GPL'd software to
stringently control access to other copyrighted works should effectively
prevent any private or public parties from invoking DMCA-like laws against
users who escape technical restriction measures implemented by GPL'd
software.
\section{GPLv3~\S4: Verbatim Copying}
\label{GPLv3s4}
GPLv3~\S4 is a revision of GPLv2\~S1 (as discussed in \S~\ref{GPLv2s1} of
this tutorial). There are almost no changes to this section from the
GPLv2\~S1, other than to use the new defined terms.
The only notable change of ``a fee'' to ``any price or no price'' in the
first sentence of GPLv3\S4\P2. The GPLv2\S1\P1 means that the GPL permits
one to charge money for the distribution of software. Despite efforts by
copyleft advocates to explain this in GPLv2 itself and in other documents,
there are evidently some people who still believe that GPLv2 allows charging
for services but not for selling copies of software and/or that the GPL
requires downloads to be gratis. Perhaps this is because GPLv2 referred to
charging a ``fee''; the term ``fee'' is generally used in connection with
services.
GPLv2's wording also referred to ``the physical act of transferring.'' The
intention was to distinguish charging for transfers from attempts to impose
licensing fees on all third parties. ``Physical'' might be read, however, as
suggesting ``distribution in a physical medium only''.
To address these two issues, GPLv3 says ``price'' in place of ``fee,'' and
removes the term ``physical.''
GPLv3~\S4 has also been revised from its corresponding section in GPLv2 in
light of the GPLv3~\S7 (see \S~\ref{GPLv3s7} in this tutorial for more).
Specifically, a distributor of verbatim copies of the program's source code
must obey any existing additional terms that apply to parts of the program
pursuant to GPLv3~\S7. In addition, the distributor is required to keep
intact all license notices, including notices of such additional terms.
Finally, there is no harm in explicitly pointing out what ought to be
obvious: that those who convey GPL-covered software may offer commercial
services for the support of that software.
\section{GPLv3~\S5: Modified Source}
\label{GPLv3s5}
GPLv3\S5 is the rewrite of GPLv2\S2, which was discussed in \S~\ref{GPLv2s2}
of this tutorial. This section discusses the changes found in GPLv3\S5
compared to GPLv2\S2.
GPLv3\S5(a) still requires modified versions be marked with ``relevant
date'', but no longer says ``the date of any change''. The best practice is
to include the date of the latest and/or most significant changes and who
made those. Of course, compared to its GPLv2\S2(a), GPLv3\S5(a) slightly
relaxes the requirements regarding notice of changes to the program. In
particular, the modified files themselves need no longer be marked. This
reduces administrative burdens for developers of modified versions of GPL'd
software.
GPLv3\S5(b) is a new but simple provision. GPLv3\S5(b) requires that the
license text itself must be unmodified (except as permitted by GPLv3\S7; see
\S~\ref{GPLv3s7} in this tutorial). Furthermore, it removes any perceived
conflict between the words ``keep intact all notices'' in GPLv3\S4, since
operating under GPLv3\S5 still includes all the requirements of GPLv3\S4 by
reference.
GPLv3\S5(c) is the primary source-code-related copyleft provision of GPL. (The
object-code-related copyleft provisions are in GPLv3\S6, discussed in
\S~\ref{GPLv3s6} of this tutorial). Compared to GPLv2\S2(b), GPLv3\S5(c)
states that the GPL applies to the whole of the work. Such was stated
already in GPLv2\S2(b), in ``in whole or in part'', but this simplified
wording makes it clear the entire covered work
Another change in GPLv3\S5(c) is the removal of the
words ``at no charge,'' which was often is misunderstood upon na\"{i}ve
reading of in GPLv2\S(b) (as discussed in \S~\ref{GPLv2s2-at-no-charge} of this
tutorial).
% FIXME-LATER: Write up something on 5d, and related it to Appropriate Legal Notices.
Note that of GPLv2~\S2's penultimate and ante-penultimate paragraphs are now
handled adequately by the definitions in GPLv3\S0 and as such, have no direct
analogs in GPLv3.
GPLv2~\S2's final paragraph, however, is reworded and expanded into the final
paragraph of GPLv3\S5, which now also covers issues related to copyright
compilations (but not compilations into object code --- that's in the next
section!). The intent and scope is the same as was intended in GPLv2.
\section{GPLv3~\S6: Non-Source and Corresponding Source}
\label{GPLv3s6}
GPLv3~\S6 clarifies and revises GPLv2~\S3. It requires distributors of GPL'd
object code to provide access to the corresponding source code, in one of
four specified ways. As noted in \S~\ref{GPLv3s0}, ``object code'' in GPLv3
is defined broadly to mean any non-source version of a work.
% FIXME: probably mostly still right, needs some updates, though.
GPLv3~\S6(a--b) now apply specifically to distribution of object code in a
physical product. Physical products include embedded systems, as well as
physical software distribution media such as CDs. As in GPLv2~\S3 (discussed
in \S~\ref{GPLv2s3} of this tutorial), the distribution of object code may
either be accompanied by the machine-readable source code, or it may be
accompanied by a valid written offer to provide the machine-readable source
code. However, unlike in GPLv2, that offer cannot be exercised by any third
party; rather, only those ``who possesses the object code'' it can exercised
the offer. (Note that this is a substantial narrowing of requirements of
offer fulfillment, and is a wonderful counterexample to dispute claims that
the GPLv3 has more requirements than GPLv2.)
% FIXME: probably mostly still right, needs some updates, though.
GPLv3~\S6(b) further revises the requirements for the written offer to
provide source code. As before, the offer must remain valid for at least
three years. In addition, even after three years, a distributor of a product
containing GPL'd object code must offer to provide source code for as long as
the distributor also continues to offer spare parts or customer support for
the product model. This is a reasonable and appropriate requirement; a
distributor should be prepared to provide source code if he or she is
prepared to provide support for other aspects of a physical product.
GPLv3~\S6(a--b) clarifies that the medium for software interchange on which
the machine-readable source code is provided must be a durable physical
medium. GPLv3~\S6(b)(2), however, permits a distributor to instead offer to
provide source code from a network server instead, which is yet another
example GPLv3 looser in its requirements than GPLv2 (see
\S~\ref{GPLv2s3-medium-customarily} for details).
% FIXME-LATER: more information about source provision, cost of physically
% performing, reasonable fees, medium customary clearly being said durable
% connecting back to previous text
GPLv3\S6(c) gives narrower permission than GPLv2\S3(c). The ``pass along''
option for GPLv3\S6(c)(1) offers is now available only for individual
distribution of object code; moreover, such individual distribution can occur
only ``occasionally and noncommercially.'' A distributor cannot comply with
the GPL merely by making object code available on a publicly-accessible
network server accompanied by a copy of the written offer to provide source
code received from an upstream distributor.
%FIXME-LATER: tie back to the discussion of the occasional offer pass along
% stuff in GPLv2 this
% FIXME: probably mostly still right, needs some updates, though.
New subsection 6d, which revises the final paragraph of GPLv2 section 3,
addresses distribution of object code by offering access to copy the code
from a designated place, such as by enabling electronic access to a network
server. Subsection 6d clarifies that the distributor must offer equivalent
access to copy the source code ``in the same way through the same place.''
This wording permits a distributor to offer a third party access to both
object code and source code on a single network portal or web page, even
though the access may include links to different physical servers. For
example, a downstream distributor may provide a link to an upstream
distributor's server and arrange with the operator of that server to keep the
source code available for copying for as long as the downstream distributor
enables access to the object code. This codifies what has been our
interpretation of GPLv2.
% FIXME: where should this go?
We improved the wording of this sentence to provide a clearer expression of
the intended policy. Under the 6d option, you may charge for the conveyed
object code. Those who pay to obtain the object code must be given equivalent
and gratis access to obtain the Corresponding Source. (If you convey the
object code to them gratis, you must likewise make the Corresponding Source
available to them without charge.) Those who do not obtain the object code
from you, perhaps because they choose not to pay the fee you charge, are
outside the scope of the provision; you need not give them any kind of access
to the Corresponding Source.
%FIXME: 6e, peer-to-peer
Informing the peers is clearly enough; what seemed to be an additional
knowledge requirement was superfluous wording.
% FIXME: Not final paragraph anymore.
The final paragraph of section 6 takes account of the fact that the Complete
Corresponding Source Code may include added parts that carry non-GPL terms,
as permitted by section 7.
% FIXME: update lock-down section to work with more recent drafts
Though the definition of Complete Corresponding Source Code in the second
paragraph of section 1 is expansive, it is not sufficient to protect users'
freedoms in many circumstances. For example, a GPL'd program, or a modified
version of such a program, might need to be signed with a key or authorized
with a code in order for it to run on a particular machine and function
properly. Similarly, a program that produces digitally-restricted files might
require a decryption code in order to read the output.
The third paragraph of section 1 addresses this problem by making clear that
Complete Corresponding Source Code includes any such encryption,
authorization, and decryption codes. By requiring the inclusion of this
information whenever the GPL requires distribution of Complete Corresponding
Source Code, we thwart efforts to obstruct the goals of the GPL, and we
ensure that users will remain in control over their own machines. We
recognize an exception where use of the program normally implies that the
user already has the codes. For example, in secure systems a computer owner
might possess any keys needed to run a program, while the distributor of the
program might not have the keys.
% FIXME: installation information??
% FIXME: perhaps this additional information isn't needed, next 3 paras, but
% there might be something good here
Another major goal for GPLv3 has been to thwart technical measures such as
signature checks in hardware to prevent modification of GPLed software on a
device. Previous drafts attempted to accomplish this by defining
"Corresponding Source" to include any encryption or authorization keys
necessary to install new versions of the software. A number of members of
the community questioned the impact and utility of such a definition.
The third discussion draft uses a different strategy to accomplish the same
task. Section 6 requires that parties distributing object code provide
recipients with the source code through certain means. Now, when those
distributors pass on the source, they are also required to pass on any
information or data necessary to install modified software on the
particular device that included it. We believe that this will more
precisely accomplish our goals, and avoid potential problems with expanding
the definition of source code. The new strategy should be familiar to free
software developers: the GNU LGPL has long had similar requirements that
enable users to link proprietary programs to modified libraries.
\label{user-product}
In addition, the scope of these requirements has been narrowed. This draft
introduces the concept of a "User Product," which includes devices that are
sold for personal, family, or household use. Distributors are only
required to provide installation information when they convey object code
in a User Product. After some discussion with committees, we discovered
that the proposals in the second discussion draft would interfere with a
number of existing business models that don't seem to be dangerous. We
believe that this compromise will achieve the greatest success in
preventing tivoization.
In brief, we condition the right to convey object code in a defined class of
``User Products,'' under certain circumstances, on providing whatever
information is required to enable a recipient to replace the object code with
a functioning modified version.
%FIXME: this really big section on user product stuff may be too much for the
% tutorial
In our earlier drafts, the requirement to provide encryption keys
applied to all acts of conveying object code, as this requirement was
part of the general definition of Corresponding Source. Section 6 of
Draft 3 now limits the applicability of the technical restrictions
provisions to object code conveyed in, with, or specifically for use in
a defined class of ``User Products.''
In our discussions with companies and governments that use specialized
or enterprise-level computer facilities, we found that sometimes these
organizations actually want their systems not to be under their own
control. Rather than agreeing to this as a concession, or bowing to
pressure, they ask for this as a preference. It is not clear that we
need to interfere, and the main problem lies elsewhere.
While imposing technical barriers to modification is wrong regardless of
circumstances, the areas where restricted devices are of the greatest
practical concern today fall within the User Product definition. Most,
if not all, technically-restricted devices running GPL-covered programs
are consumer electronics devices, and we expect that to remain true in
the near future. Moreover, the disparity in clout between the
manufacturers and these users makes it difficult for the users to reject
technical restrictions through their weak and unorganized market
power. Even if limited to User Products, as defined in Draft 3, the
provision still does the job that needs to be done. Therefore we have
decided to limit the technical restrictions provisions to User Products
in this draft.
The core of the User Product definition is a subdefinition of ``consumer
product'' taken verbatim from the Magnuson-Moss Warranty Act, a federal
consumer protection law in the United States: ``any tangible personal
property which is normally used for personal, family, or household
purposes.''\footnote{15 U.S.C.~\S\ 2301.} The United States has had
three decades of experience of liberal judicial and administrative
interpretation of this definition in a manner favorable to consumer
rights.\footnote{The Magnuson-Moss consumer product definition itself
has been influential in the United States and Canada, having been
adopted in several state and provincial consumer protection laws.} We
mean for this body of interpretation to guide interpretation of the
consumer product subdefinition in section 6, which will provide a degree
of legal certainty advantageous to device manufacturers and downstream
licensees alike. Our incorporation of such legal interpretation is in
no way intended to work a general choice of United States law for GPLv3
as a whole. The paragraph in section 6 defining ``User Product'' and
``consumer product'' contains an explicit statement to this effect,
bracketed for discussion. We will decide whether to retain this
statement in the license text after gathering comment on it.
One well-established interpretive principle under Magnuson-Moss is that
ambiguities are resolved in favor of coverage. That is, in cases where
it is not clear whether a product falls under the definition of consumer
product, the product will be treated as a consumer product.\footnote{16
C.F.R.~\S\ 700.1(a); \textit{McFadden v.~Dryvit Systems, Inc.}, 54
U.C.C.~Rep.~Serv.2d 934 (D.~Ore.~2004).} Moreover, for a given product,
``normally used'' is understood to refer to the typical use of that type
of product, rather than a particular use by a particular buyer.
Products that are commonly used for personal as well as commercial
purposes are consumer products, even if the person invoking rights is a
commercial entity intending to use the product for commercial
purposes.\footnote{16 C.F.R. \S \ 700.1(a). Numerous court decisions
interpreting Magnuson-Moss are in accord; see, e.g., \textit{Stroebner
Motors, Inc.~v.~Automobili Lamborghini S.p.A.}, 459 F.~Supp.2d 1028,
1033 (D.~Hawaii 2006).} Even a small amount of ``normal'' personal use
is enough to cause an entire product line to be treated as a consumer
product under Magnuson-Moss.\footnote{\textit{Tandy Corp.~v.~Marymac
Industries, Inc.}, 213 U.S.P.Q.~702 (S.D.~Tex.~1981). In this case, the
court concluded that TRS-80 microcomputers were consumer products, where
such computers were designed and advertised for a variety of users,
including small businesses and schools, and had only recently been
promoted for use in the home.}
We do not rely solely on the definition of consumer product, however,
because in the area of components of dwellings we consider the settled
interpretation under Magnuson-Moss underinclusive. Depending on how
such components are manufactured or sold, they may or may not be
considered Magnuson-Moss consumer products.\footnote{Building materials
that are purchased directly by a consumer from a retailer, for improving
or modifying an existing dwelling, are consumer products under
Magnuson-Moss, but building materials that are integral component parts
of the structure of a dwelling at the time that the consumer buys the
dwelling are not consumer products. 16 C.F.R.~\S\S~700.1(c)--(f);
Federal Trade Commission, Final Action Concerning Review of
Interpretations of Magnuson-Moss Warranty Act, 64 Fed.~Reg.~19,700
(April 22, 1999); see also, e.g., \textit{McFadden}, 54
U.C.C.~Rep.~Serv.2d at 934.} Therefore, we define User Products as a
superset of consumer products that also includes ``anything designed or
sold for incorporation into a dwelling.''
Although the User Products rule of Draft 3 reflects a special concern
for individual purchasers of devices, we wrote the rule to cover a
category of products, rather than categorizing users. Discrimination
against organizational users has no place in a free software license.
Moreover, a rule that applied to individual use, rather than to use of
products normally used by individuals, would have too narrow an
effect. Because of its incorporation of the liberal Magnuson-Moss
interpretation of ``consumer product,'' the User Products rule benefits
not only individual purchasers of User Products but also all
organizational purchasers of those same kinds of products, regardless of
their intended use of the products.
we have replaced the Magnuson-Moss
reference with three sentences that encapsulate the judicial and
administrative principles established over the past three decades in the
United States concerning the Magnuson-Moss consumer product definition.
First, we state that doubtful cases are resolved in favor of coverage
under the definition. Second, we indicate that the words ``normally
used'' in the consumer product definition refer to a typical or common
use of a class of product, and not the status of a particular user or
expected or actual uses by a particular user. Third, we make clear that
the existence of substantial non-consumer uses of a product does not
negate a determination that it is a consumer product, unless such
non-consumer uses represent the only significant mode of use of that
product.
It should be clear from these added sentences that it is the general
mode of use of a product that determines objectively whether or not it
is a consumer product. One could not escape the effects of the User
Products provisions by labeling what is demonstrably a consumer product
in ways that suggest it is ``for professionals,'' for example, contrary
to what some critics of Draft 3 have suggested.
We have made one additional change to the User Products provisions of
section 6. In Draft 3 we made clear that the requirement to provide
Installation Information implies no requirement to provide warranty or
support for a work that has been modified or installed on a User
Product. The Final Draft adds that there is similarly no requirement to
provide warranty or support for the User Product itself.
% FIXME: this needs integration
In Draft 3 we instead use a definition of ``Installation Information'' in
section 6 that is as simple and clear as that goal. Installation Information
is information that is ``required to install and execute modified versions of
a covered work \dots from a modified version of its Corresponding Source,''
in the same User Product for which the covered work is conveyed. We provide
guidance concerning how much information must be provided: it ``must suffice
to ensure that the continued functioning of the modified object code is in no
case prevented or interfered with solely because modification has been
made.'' For example, the information provided would be insufficient if it
enabled a modified version to run only in a disabled fashion, solely because
of the fact of modification (regardless of the actual nature of the
modification). The information need not consist of cryptographic keys;
Installation Information may be ``any methods, procedures, authorization
keys, or other information.''
%FIXME: This probably needs work to be brought into clarity with tutorial,
%next three paragarphs.
Why do distributors only have to provide Installation Information for User
Products?
Some companies effectively outsource their entire IT department to another
company. Computers and applications are installed in the company's offices,
but managed remotely by some service provider. In some of these situations,
the hardware is locked down; only the service provider has the key, and the
customers consider that to be a desirable security feature.
We think it's unfortunate that people would be willing to give up their
freedom like this. But they should be able to fend for themselves, and the
market provides plenty of alternatives to these services that would not lock
them down. As a result, we have introduced this compromise to the draft:
distributors are only required to provide Installation Information when
they're distributing the software on a User Product, where the customers'
buying power is likely to be less organized.
This is a compromise of strategy, and not our ideals. Given the environment
we live in today --- where Digital Restrictions Management is focused largely
in consumer devices, and everyone, including large companies, is becoming
increasingly worried about the effects of DRM thanks to recent developments
like the release of Microsoft's Windows Vista --- we think that the proposed
language will still provide us with enough leverage to effectively thwart
DRM. We still believe you have a fundamental right to modify the software on
all the hardware you own; the preamble explains, ``If such problems [as
locked-down hardware] arise substantially in other domains, we stand ready
to extend this provision to those domains in future versions of the GPL, as
needed to protect the freedom of users.''
The definition of Installation Information states that the information
provided ``must suffice to ensure that the continued functioning of the
modified object code is in no case prevented or interfered with solely
because modification has been made.'' We did not consider it necessary to
define ``continued functioning'' further. However, we believed it would be
appropriate to provide some additional guidance concerning the scope of
GPLv3-compliant action or inaction that distributors of
technically-restricted User Products can take with respect to a downstream
recipient who replaces the conveyed object code with a modified version. We
make clear that GPLv3 implies no obligation ``to continue to provide support
service, warranty, or updates'' for such a work.
Most technically-restricted User Products are designed to communicate across
networks. It is important for both users and network providers to know when
denial of network access to devices running modified versions becomes a GPL
violation. We settled on a rule that permits denial of access in two cases:
``when the modification itself materially and adversely affects the operation
of the network,'' and when the modification itself ``violates the rules and
protocols for communication across the network.'' The second case is
deliberately drawn in general terms. We intend it to serve as a foundation
for development of reasonable enforcement policies that respect recipients'
right to modify while recognizing the legitimate interests of network
providers.
% FIXME: This needs merged in somewhere in here
The mere fact that use of the work implies that the user \textit{has} the key
may not be enough to ensure the user's freedom in using it. The user must
also be able to read and copy the key; thus, its presence in a special
register inside the computer does not satisfy the requirement. In an
application in which the user's personal key is used to protect privacy or
limit distribution of personal data, the user clearly has the ability to read
and copy the key, which therefore is not included in the Corresponding
Source. On the other hand, if a key is generated based on the object code, or
is present in hardware, but the user cannot manipulate that key, then the key
must be provided as part of the Corresponding Source.
% FIXME: this came from Section 1 but is now mostly in Section 6
In section 1, we have tried to limit as precisely as possible the situation
in which an encryption or signing key is part of the Corresponding Source
Code of a GPL'd work. Where someone is provided a GPL'd work, he must
receive the whole of the power to use and modify the work that was available
to preceding licensors whose permissions he automatically receives. If a key
would be necessary to install a fully functional version of the GPL'd work
from source code, the user who receives the binary must receive the key along
with the source. The requirement of full functionality, which we have
illustrated with examples, is no more optional than it would be if GPL'd
software were redistributed with an additional license condition, rather than
a technical limitation, on the uses to which modified versions could be
put.\footnote{There is a clear distinction between this situation and the
situation of authenticated modules or plug-ins distributed as part of a
multi-component software system, so that instances of the software can
verify for the user the integrity of the collection. So long as the
decision about whether to run a modified version is the user's decision,
not controlled by a preceding licensor or a third party, the vendor's
authentication key would also not qualify as part of the Corresponding
Source under the language we have adopted for Draft 2.}
% FIXME: this needs the right place.
We do not object to the practice of conveying object code in a mode not
practically susceptible to modification by any party, such as code burned in
ROM or embedded in silicon. What we find ethically objectionable is the
refusal to pass on to the downstream licensee the real right to modify,
coupled with the retention of that right in the device manufacturer or some
other party. Our text has never prohibited distribution in ROM, but we have
decided to make the point explicitly, for clarity's sake. Accordingly, our
text states that the requirement to provide Installation Information ``does
not apply if neither you nor any third party retains the ability to install
modified object code on the User Product.''
%FIXME: publicly documented format. This might work as a start on that:
Our primary objective here was to ensure that the
distributor use a generally-recognized mechanism for packaging source
code.
\section{Understanding License Compatibility}
\label{license-compatibility}
% FIXME: more about license compatibility here.
A challenge that faced the Free Software community heavily through out the
early 2000s was the proliferation of incompatible Free Software licenses. Of
course, we cannot make the GPL compatible with all such licenses. GPLv3
contains provisions that are designed to reduce license incompatibility by
making it easier for developers to combine code carrying non-GPL terms with
GPL'd code.
% FIXME: connecting text
\subsection{Additional Permissions}
% FIXME: rework and fix formatting.
The GPL is a statement of permissions, some of which have conditions.
Additional terms, terms that supplement those of the GPL, may come to be
placed on, or removed from, GPL-covered code in certain common ways. We
consider those added terms ``additional permissions'' if they grant
exceptions from the conditions of the GPL, and ``additional requirements'' if
they add conditions to the basic permissions of the GPL. The treatment of
additional permissions and additional requirements under GPLv3 is necessarily
asymmetrical, because they do not raise the same ethical and interpretive
issues; in particular, additional requirements, if allowed without careful
limitation, could transform a GPL'd program into a non-free one. With these
principles in the background, section 7 answers the following questions: (1)
How do the presence of additional terms on all or part of a GPL'd program
affect users' rights? (2) When and how may a licensee add terms to code being
distributed under the GPL? (3) When may a licensee remove additional terms?
% FIXME: FSF third person, etc.
Additional permissions present the easier case. We have licensed some of our
own software under GPLv2 with permissive exceptions that allow combination
with non-free code, and that allow removal of those permissions by downstream
recipients; similarly, LGPLv2.1 is in essence a permissive variant of GPLv2,
and it permits relicensing under the GPL. We have generalized these
practices in section 7. A licensee may remove any additional permission from
a covered work, whether it was placed by the original author or by an
upstream distributor. A licensee may also add any kind of additional
permission to any part of a work for which the licensee has, or can give,
appropriate copyright permission. For example, if the licensee has written
that part, the licensee is the copyright holder for that part and can
therefore give additional permissions that are applicable to it.
Alternatively, the part may have been written by someone else and licensed,
with the additional permissions, to that licensee. Any additional
permissions on that part are, in turn, removable by downstream recipients.
As subsection 7a explains, the effect of an additional permission depends on
whether the permission applies to the whole work or a part.
% FIXME: rework this a bit
We have drafted version 3 of the GNU LGPL, which we have released with Draft
2 of GPLv3, as a simple list of additional permissions supplementing the
terms of GPLv3. Section 7 has thus provided the basis for recasting a
formally complex license as an elegant set of added terms, without changing
any of the fundamental features of the existing LGPL. We offer this draft of
LGPLv3 as as a model for developers wishing to license their works under the
GPL with permissive exceptions. The removability of additional permissions
under section 7 does not alter any existing behavior of the LGPL; the LGPL
has always allowed relicensing under the ordinary GPL.
\subsection{Additional Requirements and License Compatibility}
% FIXME: minor rewrites needed
We broadened the title of section 7 because license compatibility, as it is
conventionally understood, is only one of several facets of the placement of
additional terms on GPL'd code. The license compatibility issue arises for
three reasons. First, the GPL is a strong copyleft license, requiring
modified versions to be distributed under the GPL. Second, the GPL states
that no further restrictions may be placed on the rights of recipients.
Third, all other free software licenses in common use contain certain
requirements, many of which are not conditions made by the GPL. Thus, when
GPL'd code is modified by combination with code covered by another formal
license that specifies other requirements, and that modified code is then
distributed to others, the freedom of recipients may be burdened by
additional requirements in violation of the GPL. It can be seen that
additional permissions in other licenses do not raise any problems of license
compatibility.
% FIXME: minor rewrites needed
Section 7 relaxes the prohibition on further restrictions slightly by
enumerating, in subsection 7b, a limited list of categories of additional
requirements that may be placed on code without violating GPLv3. The list
includes the items that were listed in Draft 1, though rewritten for clarity.
It also includes a new catchall category for terms that might not obviously
fall within one of the other categories but which are precisely equivalent to
GPLv3 conditions, or which deny permission for activities clearly not
permitted by GPLv3. We have carefully considered but rejected proposals to
expand this list further. We have also rejected suggestions, made by some
discussion committee members, that the Affero clause requirement (7d in Draft
1 and 7b4 in Draft 2) be removed, though we have revised it in response to
certain comments. We are unwavering in our view that the Affero requirement
is a legitimate one, and we are committed to achieving compatibility of the
Affero GPL with GPLv3.
% FIXME: minor rewrites needed
A GPL licensee may place an additional requirement on code for which the
licensee has or can give appropriate copyright permission, but only if that
requirement falls within the list given in subsection 7b. Placement of any
other kind of additional requirement continues to be a violation of the
license. Additional requirements that are in the 7b list may not be removed,
but if a user receives GPL'd code that purports to include an additional
requirement not in the 7b list, the user may remove that requirement. Here
we were particularly concerned to address the problem of program authors who
purport to license their works in a misleading and possibly
self-contradictory fashion, using the GPL together with unacceptable added
restrictions that would make those works non-free software.
\section{GPLv3~\S7: Explicit Compatibility}
% FIXME: probably mostly still right, needs some updates, though.
In GPLv3 we take a new approach to the issue of combining GPL'd code with
code governed by the terms of other free software licenses. Our view, though
it was not explicitly stated in GPLv2 itself, was that GPLv2 allowed such
combinations only if the non-GPL licensing terms permitted distribution under
the GPL and imposed no restrictions on the code that were not also imposed by
the GPL. In practice, we supplemented this policy with a structure of
exceptions for certain kinds of combinations.
% FIXME: probably mostly still right, needs some updates, though.
Section 7 of GPLv3 implements a more explicit policy on license
compatibility. It formalizes the circumstances under which a licensee may
release a covered work that includes an added part carrying non-GPL terms. We
distinguish between terms that provide additional permissions, and terms that
place additional requirements on the code, relative to the permissions and
requirements established by applying the GPL to the code.
% FIXME: probably mostly still right, needs some updates, though.
Section 7 first explicitly allows added parts covered by terms with
additional permissions to be combined with GPL'd code. This codifies our
existing practice of regarding such licensing terms as compatible with the
GPL. A downstream user of a combined GPL'd work who modifies such an added
part may remove the additional permissions, in which case the broader
permissions no longer apply to the modified version, and only the terms of
the GPL apply to it.
% FIXME: probably mostly still right, needs some updates, though.
In its treatment of terms that impose additional requirements, section 7
extends the range of licensing terms with which the GPL is compatible. An
added part carrying additional requirements may be combined with GPL'd code,
but only if those requirements belong to an set enumerated in section 7. We
must, of course, place some limit on the kinds of additional requirements
that we will accept, to ensure that enhanced license compatibility does not
defeat the broader freedoms advanced by the GPL. Unlike terms that grant
additional permissions, terms that impose additional requirements cannot be
removed by a downstream user of the combined GPL'd work, because no such user
would have the right to do so.
% FIXME: probably mostly still right, needs some updates, though.
Under subsections 7a and 7b, the requirements may include preservation of
copyright notices, information about the origins of the code or alterations
of the code, and different warranty disclaimers. Under subsection 7c, the
requirements may include limitations on the use of names of contributors and
on the use of trademarks for publicity purposes. In general, we permit these
requirements in added terms because many free software licenses include them
and we consider them to be unobjectionable. Because we support trademark fair
use, the limitations on the use of trademarks may seek to enforce only what
is required by trademark law, and may not prohibit what would constitute fair
use.
% FIXME: 7d-f
\section{GPLv3~\S7(e): Peer-to-Peer Sharing Networks}
% FIXME: rewrite a bit, maybe drop reference to bitorrent?
Certain decentralized forms of peer-to-peer file sharing present a challenge
to the unidirectional view of distribution that is implicit in GPLv2 and
Draft 1 of GPLv3. It is neither straightforward nor reasonable to identify
an upstream/downstream link in BitTorrent distribution; such distribution is
multidirectional, cooperative and anonymous. In systems like BitTorrent,
participants act both as transmitters and recipients of blocks of a
particular file, but they see themselves as users and receivers, and not as
distributors in any conventional sense. At any given moment of time, most
peers will not have the complete file.
% FIXME: rewrite a bit.
The GPL permits distribution of a work in object code form over a network,
provided that the distributor offers equivalent access to copy the
Corresponding Source Code ``in the same way through the same place.'' This
wording might be interpreted to permit BitTorrent distribution of binaries if
they are packaged together with the source code, but this impractical, for at
least two reasons. First, even if the source code is packaged with the
binary, it will only be available to a non-seeding peer at the end of the
distribution process, but the peer will already have been providing parts of
the binary to others in the network, functioning rather like a router or a
cache proxy. Second, in practice BitTorrent and similar peer-to-peer forms
of transmission have been less suitable means for distributing source code.
In large distributions, packaging source code with the binary may result in a
substantial increase in file size and transmission time. Source code
packages themselves tend not to be transmitted through BitTorrent owing to
reduced demand. There generally will be too few participants downloading the
same source package at the same time to enable effective seeding and
distribution.
% FIXME: rewrite a bit.
We have made two changes that recognize and facilitate distribution of
covered works in object code form using BitTorrent or similar peer-to-peer
methods. First, under new subsection 6e, if a licensee conveys such a work
using peer-to-peer transmission, that licensee is in compliance with section
6 so long as the licensee knows, and informs other peers where, the object
code and its Corresponding Source are publicly available at no charge under
subsection 6d. The Corresponding Source therefore need not be provided
through the peer-to-peer system that was used for providing the binary.
Second, we have revised section 9 to make clear that ancillary propagation of
a covered work that occurs as part of the process of peer-to-peer file
transmission does not require acceptance, just as mere receipt and execution
of the Program does not require acceptance. Such ancillary propagation is
permitted without limitation or further obligation.
% FIXME: removing additional restrictions
% FIXME: probably mostly still right, needs some updates, though.
Section 7 requires a downstream user of a covered work to preserve the
non-GPL terms covering the added parts just as they must preserve the GPL, as
long as any substantial portion of those parts is present in the user's
version.
% FIXME: minor rewrites needed
Section 7 points out that GPLv3 itself makes no assertion that an additional
requirement is enforceable by the copyright holder. However, section 7 makes
clear that enforcement of such requirements is expected to be by the
termination procedure given in section 8 of GPLv3.
% FIXME: better context, etc.
Some have questioned whether section 7 is needed, and some have suggested
that it creates complexity that did not previously exist. We point out to
those readers that there is already GPLv2-licensed code that carries
additional terms. One of the objectives of section 7 is to rationalize
existing practices of program authors and modifiers by setting clear
guidelines regarding the removal and addition of such terms. With its
carefully limited list of allowed additional requirements, section 7
accomplishes additional objectives, permitting the expansion of the base of
code available for GPL developers, while also encouraging useful
experimentation with requirements we do not include in the GPL itself.
\section{GPLv3~\S8: A Lighter Termination}
% FIXME: probably mostly still right, needs some updates, though.
GPLv2 provided for automatic termination of the rights of a person who
copied, modified, sublicensed, or distributed a work in violation of the
license. Automatic termination can be too harsh for those who have committed
an inadvertent violation, particularly in cases involving distribution of
large collections of software having numerous copyright holders. A violator
who resumes compliance with GPLv2 would need to obtain forgiveness from all
copyright holders, but even to contact them all might be impossible.
% FIXME: needs to be updated to describe more complex termination
Section 8 of GPLv3 replaces automatic termination with a non-automatic
termination process. Any copyright holder for the licensed work may opt to
terminate the rights of a violator of the license, provided that the
copyright holder has first given notice of the violation within 60 days of
its most recent occurrence. A violator who has been given notice may make
efforts to enter into compliance and may request that the copyright holder
agree not exercise the right of termination; the copyright holder may choose
to grant or refuse this request.
% FIXME: needs to be updated to describe more complex termination
If a licensee who is in violation of GPLv3 acts to correct the violation and
enter into compliance, and the licensee receives no notice of the past
violation within 60 days, then the licensee need not worry about termination
of rights under the license.
In Draft 3 the termination provision of section 8 has been revised to
indicate that, if a licensee violates the GPL, a contributor may terminate
any patent licenses that it granted under the first paragraph of section 11
to that licensee, in addition to any copyright permissions the contributor
granted to the licensee. Therefore, a contributor may terminate the patent
licenses it granted to a downstream licensee who brings patent infringement
litigation in violation of section 10.
We have made two substantive changes to section 8. First, we have clarified
that patent rights granted under the GPL are among the rights that a
copyright holder may terminate under section 8. Therefore, a contributor who
grants a patent license under the first paragraph of section 11 may terminate
that patent license, just as that contributor may terminate copyright rights,
to a downstream recipient who has violated the license. We think that this
is a reasonable result, and was already implicit in the wording of the
termination provision in our earlier drafts. Moreover, this clarification
should encourage patent holders to make contributions to GPL-covered
programs.
Second, we have modified the termination procedure by providing a limited
opportunity to cure license violations, an improvement that was requested by
many different members of our community. If a licensee has committed a
first-time violation of the GPL with respect to a given copyright holder, but
the licensee cures the violation within 30 days following receipt of notice
of the violation, then any of the licensee's GPL rights that have been
terminated by the copyright holder are ``automatically reinstated.'' The
addition of the cure opportunity achieves a better balance than our earlier
section 8 drafts between facilitating enforcement of the license and
protecting inadvertent violators against unfair results.
We have restructured the form of section 8 by replacing non-automatic
termination with automatic termination coupled with opportunities for
provisional and permanent reinstatement of rights. The revised wording does
not alter the underlying policy or details of procedure established in the
previous drafts, including the 60-day period of repose and 30-day cure
opportunity for first-time violators. The restoration of automatic
termination was motivated in part to facilitate enforcement in European
countries. We also believe the revised wording will be easier to understand
and apply in all jurisdictions.
\section{GPLv3~\S9: Acceptance}
% FIXME: needs some work here
Section 9 means what it says: mere receipt or execution of code neither
requires nor signifies contractual acceptance under the GPL. Speaking more
broadly, we have intentionally structured our license as a unilateral grant
of copyright permissions, the basic operation of which exists outside of any
law of contract. Whether and when a contractual relationship is formed
between licensor and licensee under local law do not necessarily matter to
the working of the license.
\section{GPLv3~\S10: Explicit Downstream License}
% FIXME: These don't belong here, but it's closer to where it ought to be now.
It is important to note that section 11, paragraph 3 refers to a work that is
conveyed, and section 10, paragraph 2 refers to a kind of automatic
counterpart to conveying achieved as the result of a transaction.
% FIXME: needs filled out and more here.
Draft1 removed the words ``at no charge'' from what is now subsection 5b, the
core copyleft provision, for reasons related to our current changes to the
second paragraph of section 4: it had contributed to a misconception that the
GPL did not permit charging for distribution of copies. The purpose of the
``at no charge'' wording was to prevent attempts to collect royalties from
third parties. The removal of these words created the danger that the
imposition of licensing fees would no longer be seen as a license
violation.
We therefore have added a new explicit prohibition on imposition of licensing
fees or royalties in section 10. This section is an appropriate place for
such a clause, since it is a specific consequence of the general requirement
that no further restrictions be imposed on downstream recipients of
GPL-covered code.
Careful readers of the GPL have suggested that its explicit prohibition
against imposition of further restrictions\footnote{GPLv2, section 6; Draft
3, section 10, third paragraph.} has, or ought to have, implications for
those who assert patents against other licensees. Draft 2 took some steps to
clarify this point in a manner not specific to patents, by describing the
imposition of ``a license fee, royalty, or other charge'' for exercising GPL
rights as one example of an impermissible further restriction. In Draft 3 we
have clarified further that the requirement of non-imposition of further
restrictions has specific consequences for litigation accusing GPL-covered
programs of infringement. Section 10 now states that ``you may not initiate
litigation (including a cross-claim or counterclaim in a lawsuit) alleging
that any patent claim is infringed by making, using, selling, offering for
sale, or importing the Program (or the contribution of any contributor).''
That is to say, a patent holder's licensed permissions to use a work under
GPLv3 may be terminated under section 8 if the patent holder files a lawsuit
alleging that use of the work, or of any upstream GPLv3-licensed work on
which the work is based, infringes a patent.
\section{GPLv3~\S11: Explicit Patent Licensing}
\label{GPLv3s11}
The patent licensing practices that section 7 of GPLv2 (corresponding to
section 12 of GPLv3) was designed to prevent are one of several ways in which
software patents threaten to make free programs non-free and to prevent users
from exercising their rights under the GPL. GPLv3 takes a more comprehensive
approach to combatting the danger of patents.
Software patenting is a harmful and unjust policy, and should be abolished;
recent experience makes this all the more evident. Since many countries grant
patents that can apply to and prohibit software packages, in various guises
and to varying degrees, we seek to protect the users of GPL-covered programs
from those patents, while at the same time making it feasible for patent
holders to contribute to and distribute GPL-covered programs as long as they
do not attack the users of those programs.
It is generally understood that GPLv2 implies some limits on a licensee's
power to assert patent claims against the use of GPL-covered works.
Therefore, we have designed GPLv3 to reduce the patent risks that distort and
threaten the activities of users who make, run, modify and share free
software. At the same time, we have given due consideration to practical
goals such as certainty and administrability for patent holders that
participate in distribution and development of GPL-covered software. Our
policy requires each such patent holder to provide appropriate levels of
patent assurance to users, according to the nature of the patent holder's
relationship to the program.
Draft 3 features several significant changes concerning patents. We have
made improvements to earlier wording, clarified when patent assertion becomes
a prohibited restriction on GPL rights, and replaced a distribution-triggered
non-assertion covenant with a contribution-based patent license grant. We
have also added provisions to block collusion by patent holders with software
distributors that would extend patent licenses in a discriminatory way.
Draft 3 introduces the terms ``contributor'' and ``contribution,'' which are
used in the third paragraph of section 10 and the first paragraph of section
11, discussed successively in the following two subsections. Section 0
defines a contributor as ``a party who licenses under this License a work on
which the Program is based.'' That work is the ``contribution'' of that
contributor. In other words, each received GPLv3-covered work is associated
with one or more contributors, making up the finite set of upstream GPLv3
licensors for that work. Viewed from the perspective of a recipient of the
Program, contributors include all the copyright holders for the Program,
other than copyright holders of material originally licensed under non-GPL
terms and later incorporated into a GPL-covered work. The contributors are
therefore the initial GPLv3 licensors of the Program and all subsequent
upstream licensors who convey, under the terms of section 5, modified works
on which the Program is based.
For a contributor whose contribution is a modified work conveyed under
section 5, the contribution is ``the entire work, as a whole'' which the
contributor is required to license under GPLv3. The contribution therefore
includes not just the material added or altered by the contributor, but also
the pre-existing material the contributor copied from the upstream version
and retained in the modified version. Our usage of ``contributor'' and
``contribution'' should not be confused with the various other ways in which
those terms are used in certain other free software licenses.\footnote{Cf.,
e.g., Apache License, version 2.0, section 1; Eclipse Public License,
version 1.0, section 1; Mozilla Public License, version 1.1, section 1.1.}
The term ``patent license,'' as used in the third through fifth
paragraphs of section 11, is not meant to be confined to agreements
formally identified or classified as patent licenses. The new second
paragraph of section 11 makes this clear by defining ``patent license,''
for purposes of the subsequent three paragraphs, as ``a patent license,
a covenant not to bring suit for patent infringement, or any other
express agreement or commitment, however denominated, not to enforce a
patent.'' The definition does not include patent licenses that arise by
implication or operation of law, because the third through fifth
paragraphs of section 11 are specifically concerned with explicit
promises that purport to be legally enforceable.
Our previous drafts featured a patent license grant triggered by all
acts of distribution of GPLv3-covered works.\footnote{In Draft 2 we
rewrote the patent license as a covenant not to assert patent claims. We
explain why we reverted to the form of a patent license grant in \S\
\ref{cov}.} Many patent-holding companies objected to this policy. They
have made two objections: (1) the far-reaching impact of the patent
license grant on the patent holder is disproportionate to the act of
merely distributing code without modification or transformation, and (2)
it is unreasonable to expect an owner of vast patent assets to exercise
requisite diligence in reviewing all the GPL-covered software that it
provides to others. Some expressed particular concern about the
consequences of ``inadvertent'' distribution.
The argument that the impact of the patent license grant would be
``disproportionate,'' that is to say unfair, is not valid. Since
software patents are weapons that no one should have, and using them for
aggression against free software developers is an egregious act,
preventing that act cannot be unfair.
However, the second argument seems valid in a practical sense. A
typical GNU/Linux distribution includes thousands of programs. It would
be quite difficult for a redistributor with a large patent portfolio to
review all those programs against that portfolio every time it receives
and passes on a new version of the distribution. Moreover, this question
raises a strategic issue. If the GPLv3 patent license requirements
convince patent-holding companies to remain outside the distribution
path of all GPL-covered software, then these requirements, no matter how
strong, will cover few patents.
We concluded it would be more effective to make a partial concession
which would lead these companies to feel secure in doing the
distribution themselves, so that the conditions of section 10 would
apply to assertion of their patents. We therefore made the stricter
section 11 patent license apply only to those distributors that have
modified the program. The other changes we have made in sections 10 and
11 provide strengthened defenses against patent assertion and compensate
partly for this concession.
Therefore, in Draft 3, the first paragraph of section 11 states that a
contributor's patent license covers all the essential patent claims
implemented by the whole program as that contributor distributes it.
Contributors of modified works grant a patent license to claims that
read on ``the entire work, as a whole.'' This is the work that the
copyleft clause in section 5 requires the contributor to license under
GPLv3; it includes the material the contributor has copied from the
upstream version that the contributor has modified. The first paragraph
of section 11 does not apply to those that redistribute the program
without change.\footnote{An implied patent license from the distributor,
however, may arise by operation of law. See the final paragraph of
section 11. Moreover, distributors are subject to the limits on patent
assertion contained in the third paragraph of section 10.}
We hope that this decision will result in fairly frequent licensing of
patent claims by contributors. A contributor is charged with awareness
of the fact that it has modified a work and provided it to others; no
act of contribution should be treated as inadvertent. Our rule also
requires no more work, for a contributor, than the weaker rule proposed
by the patent holders. Under their rule, the contributor must always
compare the entire work against its patent portfolio to determine
whether the combination of the modifications with the remainder of the
work cause it to read on any of the contributor's patent claims.
We have made three changes to the definition of ``essential patent
claims'' in section 0. This definition now serves exclusively to
identify the set of patent claims licensed by a contributor under the
first paragraph of section 11.
First, we have clarified when essential patent claims include
sublicensable claims that have been licensed to the contributor by a
third party.\footnote{This issue is typically handled in other free
software licenses having patent licensing provisions by use of the
unhelpful term ``licensable,'' which is either left undefined or is
given an ambiguous definition.} Most commercial patent license
agreements that permit sublicensing do so under restrictive terms that
are inconsistent with the requirements of the GPL. For example, some
patent licenses allow the patent licensee to sublicense but require
collection of royalties from any sublicensees. The patent licensee
could not distribute a GPL-covered program and grant the recipient a
patent sublicense for the program without violating section 12 of
GPLv3.\footnote{Draft 3 provides a new example in section 12 that makes
this point clear.} In rare cases, however, a conveying party can freely
grant patent sublicenses to downstream recipients without violating the
GPL.
Draft 3 now defines essential patent claims, for a given party, as a
subset of the claims ``owned or controlled'' by the party. The
definition states that ``control includes the right to grant sublicenses
in a manner consistent with the requirements of this License.''
Therefore, in the case of a patent license that requires collection of
royalties from sublicensees, essential patent claims would not include
any claims sublicensable under that patent license, because sublicenses
to those claims could not be granted consistent with section 12.
Second, we now state that essential patent claims are those ``that would
be infringed by some manner, permitted by this License, of making,
using, or selling the work.'' This modified wording is intended to make
clear that a patent claim is ``essential'' if some mode of usage would
infringe that claim, even if there are other modes of usage that would
not infringe.
Third, we have clarified that essential patent claims ``do not include
claims that would be infringed only as a consequence of further
modification of the work.'' That is to say, the set of essential patent
claims licensed under the first paragraph of section 11 is fixed by the
the particular version of the work that was contributed. The claim set
cannot expand as a work is further modified downstream. (If it could,
then any software patent claim would be included, since any software
patent claim can be infringed by some further modification of the
work.)\footnote{However, ``the work'' should not be understood to be
restricted to a particular mechanical affixation of, or medium for
distributing, a program, where the same program might be provided in
other forms or in other ways that may be captured by other patent claims
held by the contributor.}
The downstream shielding provision of section 11 responds particularly
to the problem of exclusive deals between patent holders and
distributors, which threaten to distort the free software distribution
system in a manner adverse to developers and users. Draft 2 added a
source code availability option to this provision, as a specific
alternative to the general requirement to shield downstream users from
patent claims licensed to the distributor. A distributor conveying a
covered work knowingly relying on a patent license may comply with the
provision by ensuring that the Corresponding Source of the work is
publicly available, free of charge. We retained the shielding option in
Draft 2 because we did not wish to impose a general requirement to make
source code available to all, which has never been a GPL condition.
The addition of the source code availability option was supported by the
free software vendors most likely to be affected by the downstream
shielding provision. Enterprises that primarily use and occasionally
distribute free software, however, raised concerns regarding the
continued inclusion of a broadly-worded requirement to ``shield,'' which
appears to have been mistakenly read by those parties as creating an
obligation to indemnify. To satisfy these concerns, in Draft 3 we have
replaced the option to shield with two specific alternatives to the
source code availability option. The distributor may comply by
disclaiming the patent license it has been granted for the conveyed
work, or by arranging to extend the patent license to downstream
recipients.\footnote{The latter option, if chosen, must be done ``in a
manner consistent with the requirements of this License''; for example,
it is unavailable if extension of the patent license would result in a
violation of section 12. Cf.~the discussion of sublicensable patent
claims in \S\ \ref{epc}.} The GPL is intended to permit private
distribution as well as public distribution, and the addition of these
options ensures that this remains the case, even though we expect that
distributors in this situation will usually choose the source code
availability option.
Without altering its underlying logic, we have modified the phrasing of
the requirement to make clear that it is activated only if the
Corresponding Source is not already otherwise publicly available. (Most
often it will, in fact, already be available on some network server
operated by a third party.) Even if it is not already available, the
option to ``cause the Corresponding Source to be so available'' can then
be satisfied by verifying that a third party has acted to make it
available. That is to say, the affected distributor need not itself
host the Corresponding Source to take advantage of the source code
availability option. This subtlety may help the distributor avoid
certain peculiar assumptions of liability.
We have made two other changes to the downstream shielding provision.
The phrase ``knowingly rely'' was left undefined in our earlier drafts;
in Draft 3 we have provided a detailed definition. We have also deleted
the condition precedent, added in Draft 2, that the relied-upon patent
license be one that is non-sublicensable and ``not generally available
to all''; this was imprecise in Draft 2 and is unnecessary in Draft
3. In nearly all cases in which the ``knowingly relying'' test is met,
the patent license will indeed not be sublicensable or generally
available to all on free terms. If, on the other hand, the patent
license is generally available under terms consistent with the
requirements of the GPL, the distributor is automatically in compliance,
because the patent license has already been extended to all downstream
recipients. If the patent license is sublicensable on GPL-consistent
terms, the distributor may choose to grant sublicenses to downstream
recipients instead of causing source code to be publicly available. In
such a case, if the distributor is also a contributor, it will already
have granted a patent sublicense by operation of the first paragraph of
section 11,\footnote{See \S\ \ref{epc}.} and so it need not do anything
further to comply with the third paragraph.
% FIXME: This probably needs editing
One major goal for GPLv3 is to provide developers with additional protection
from being sued for patent infringement. After much feedback and cooperation
from the committees, we are now proposing a patent license which closely
resembles those found in other free software licenses. This will be more
comfortable for everyone in the free software community to use, without
creating undue burdens for distributors.
We have also added new terms to stop distributors from colluding with third
parties to offer selective patent protection, as Microsoft and Novell have
recently done. The GPL is designed to ensure that all users receive the
same rights; arrangements that circumvent this make a mockery of free
software, and we must do everything in our power to stop them.
Our strategy has two parts. First, any license that protects some
recipients of GPLed software must be extended to all recipients of the
software. Second, we prohibit anyone who made such an agreement from
distributing software released under GPLv3. We are still considering
whether or not this ban should apply when a deal was made before these
terms were written, and we look forward to community input on this issue.
The patent license grant of the first paragraph of section 11 no longer
applies to those who merely distribute works without modification. (We
explain why we made this change in the next subsection.) Such parties are
nonetheless subject to the conditions stated in section 10. Unlike the
patent license, which establishes a defense for downstream users lasting for
as long as they remain in compliance with the GPL, the commitment not to sue
that arises under section 10 is one that the distributor can end, so long as
the distributor also ceases to distribute. This is because a party who
initiates patent litigation in violation of section 10 risks termination of
its licensed permissions by the copyright holders of the work.
% FIXME: just brought in words here, needs rewriting.
is rooted in the basic principles of the GPL.
Our license has always stated that distributors may not impose further
restrictions on users' exercise of GPL rights. To make the suggested
distinction between contribution and distribution is to allow a
distributor to demand patent royalties from a direct or indirect
recipient, based on claims embodied in the distributed code. This
undeniably burdens users with an additional legal restriction on their
rights, in violation of the license.
%FIXME: possible useful text, but maybe not.
In the covenant provided in the revised section 11, the set of claims
that a party undertakes not to assert against downstream users are that
party's ``essential patent claims'' in the work conveyed by the party.
``Essential patent claims,'' a new term defined in section 0, are simply
all claims ``that would be infringed by making, using, or selling the
work.'' We have abandoned the phrase ``reasonably contemplated use.''
This change makes the obligations of distributing patent holders more
predictable.
% FIXME: probably needs a lot of work, these provisions changed over time.
GPLv3 adds a new section on licensing of patents. GPLv2 relies on an implied
patent license. The doctrine of implied license is one that is recognized
under United States patent law but may not be recognized in other
jurisdictions. We have therefore decided to make the patent license grant
explicit in GPLv3. Under section 11, a redistributor of a GPL'd work
automatically grants a nonexclusive, royalty-free and worldwide license for
any patent claims held by the redistributor, if those claims would be
infringed by the work or a reasonably contemplated use of the work.
% FIXME: probably needs a lot of work, these provisions changed over time.
The patent license is granted both to recipients of the redistributed work
and to any other users who have received any version of the work. Section 11
therefore ensures that downstream users of GPL'd code and works derived from
GPL'd code are protected from the threat of patent infringement allegations
made by upstream distributors, regardless of which country's laws are held to
apply to any particular aspect of the distribution or licensing of the GPL'd
code.
% FIXME: probably needs a lot of work, these provisions changed over time.
A redistributor of GPL'd code may benefit from a patent license that has been
granted by a third party, where the third party otherwise could bring a
patent infringement lawsuit against the redistributor based on the
distribution or other use of the code. In such a case, downstream users of
the redistributed code generally remain vulnerable to the applicable patent
claims of the third party. This threatens to defeat the purposes of the GPL,
for the third party could prevent any downstream users from exercising the
freedoms that the license seeks to guarantee.
% FIXME: probably needs a lot of work, these provisions changed over time.
The second paragraph of section 11 addresses this problem by requiring the
redistributor to act to shield downstream users from these patent claims. The
requirement applies only to those redistributors who distribute knowingly
relying on a patent license. Many companies enter into blanket patent
cross-licensing agreements. With respect to some such agreements, it would
not be reasonable to expect a company to know that a particular patent
license covered by the agreement, but not specifically mentioned in it,
protects the company's distribution of GPL'd code.
% FIXME: does this still fit with the final retaliation provision?
This narrowly-targeted patent retaliation provision is the only form of
patent retaliation that GPLv3 imposes by its own force. We believe that it
strikes a proper balance between preserving the freedom of a user to run and
modify a program, and protecting the rights of other users to run, modify,
copy, and distribute code free from threats by patent holders. It is
particularly intended to discourage a GPL licensee from securing a patent
directed to unreleased modifications of GPL'd code and then suing the
original developers or others for making their own equivalent modifications.
Several other free software licenses include significantly broader patent
retaliation provisions. In our view, too little is known about the
consequences of these forms of patent retaliation. As we explain below,
section 7 permits distribution of a GPL'd work that includes added parts
covered by terms other than those of the GPL. Such terms may include certain
kinds of patent retaliation provisions that are broader than those of section
2.
% FIXME: should we mention Microsoft-Novell at all?
Section 7 of GPLv2 (now section 12 of GPLv3) has seen some success in
deterring conduct that would otherwise result in denial of full downstream
enjoyment of GPL rights. Experience has shown us that more is necessary,
however, to ensure adequate community safety where companies act in concert
to heighten the anticompetitive use of patents that they hold or license.
Previous drafts of GPLv3 included a ``downstream shielding'' provision in
section 11, which we have further refined in Draft 3; it is now found in the
third paragraph of section 11. In addition, Draft 3 introduces two new
provisions in section 11, located in the fourth and fifth paragraphs, that
address the problem of collusive extension of patent forbearance promises
that discriminate against particular classes of users and against the
exercise of particular freedoms. This problem has been made more acute by the
recent Microsoft/Novell deal.
We attack the Microsoft-Novell deal from two angles. First, in the sixth
paragraph of section 11, the draft says that if you arrange to provide patent
protection to some of the people who get the software from you, that
protection is automatically extended to everyone who receives the software,
no matter how they get it. This means that the patent protection Microsoft
has extended to Novell's customers would be extended to everyone who uses any
software Novell distributes under GPLv3.
Second, in the seventh paragraph, the draft says that you are prohibited from
distributing software under GPLv3 if you make an agreement like the
Microsoft-Novell deal in the future. This will prevent other distributors
from trying to make other deals like it.
The main reason for this is tactical. We believe we can do more to
protect the community by allowing Novell to use software under GPL
version 3 than by forbidding it to do so. This is because of
paragraph 6 of section 11 (corresponding to paragraph 4 in Draft 3).
It will apply, under the Microsoft/Novell deal, because of the coupons
that Microsoft has acquired that essentially commit it to participate
in the distribution of the Novell SLES GNU/Linux system.
Microsoft is scrambling to dispose of as many Novell SLES coupons as
possible prior to the adoption of GPLv3. Unfortunately for Microsoft,
those coupons bear no expiration date, and paragraph 6 has no cut-off
date. Through its ongoing distribution of coupons, Microsoft will
have procured the distribution of GPLv3-covered programs as soon as
they are included in Novell SLES distributions, thereby extending
patent defenses to all downstream recipients of that software by
operation of paragraph 6.
A secondary reason is to avoid affecting other kinds of agreements for
other kinds of activities. We have tried to take care in paragraph 7
to distinguish pernicious deals of the Microsoft/Novell type from
business conduct that is not particularly harmful, but we cannot be
sure we have entirely succeeded. There remains some risk that other
unchangeable past agreements could fall within its scope.
In future deals, distributors engaging in ordinary business practices
can structure the agreements so that they do not fall under paragraph
7. However, it will block Microsoft and other patent aggressors from
further such attempts to subvert parts of our community.
A software patent forbids the use of a technique or algorithm, and its
existence is a threat to all software developers and users. A patent
holder can use a patent to suppress any program which implements the
patented technique, even if thousands of other techniques are
implemented together with it. Both free software and proprietary
software are threatened with death in this way.
However, patents threaten free software with a fate worse than death: a
patent holder might also try to use the patent to impose restrictions on
use or distribution of a free program, such as to make users feel they
must pay for permission to use it. This would effectively make it
proprietary software, exactly what the GPL is intended to prevent.
Novell and Microsoft have recently attempted a new way of using patents
against our community, which involves a narrow and discriminatory
promise by a patent holder not to sue customers of one particular
distributor of a GPL-covered program. Such deals threaten our community
in several ways, each of which may be regarded as de facto
proprietization of the software. If users are frightened into paying
that one distributor just to be safe from lawsuits, in effect they are
paying for permission to use the program. They effectively deny even
these customers the full and safe exercise of some of the freedoms
granted by the GPL. And they make disfavored free software developers
and distributors more vulnerable to attacks of patent aggression, by
dividing them from another part of our community, the commercial users
that might otherwise come to their defense.
We have added the fourth and fifth paragraphs of section 11 to combat
this threat. This subsection briefly describes the operation of the new
provisions. We follow it with a more detailed separate note on the
Microsoft/Novell patent deal, in which we provide an extensive rationale
for these measures.
As noted, one effect of the discriminatory patent promise is to divide
and isolate those who make free software from the commercial users to
whom the promise is extended. This deprives the noncommercial
developers of the communal defensive measures against patents made
possible by the support of those commercial users. The fourth paragraph
of section 11 operates to restore effective defenses to the targets of
patent aggression.
A patent holder becomes subject to the fourth paragraph of section 11
when it enters into a transaction or arrangement that involves two acts:
(1) conveying a GPLv3-covered work, and (2) offering to some, but not
all, of the work's eventual users a patent license for particular
activities using specific copies of the covered work. This paragraph
only operates when the two triggering acts are part of a single
arrangement, because the patent license is part of the arrangement for
conveying, which requires copyright permission. Under those conditions,
the discriminatory patent license is ``automatically extended to all
recipients of the covered work and works based on it.''
This provision establishes a defense to infringement allegations brought
by the patent holder against any users of the program who are not
covered by the discriminatory patent license. That is to say, it gives
all recipients the benefit of the patent promise that the patent holder
extended only to some. The effect is to make contributing discriminatory
promises of patent safety to a GPL distribution essentially like
contributing code. In both cases, the operation of the GPL extends
license permission to everyone that receives a copy of the program.
The fourth paragraph of section 11 gives users a defense against patent
aggression brought by the party who made the discriminatory patent
promise that excluded them. By contrast, the fifth paragraph stops free
software vendors from contracting with patent holders to make
discriminatory patent promises. In effect, the fifth paragraph extends
the principle of section 12 to situations involving collusion between a
patent holder and a distributor.
Under this provision, a distributor conveying a GPL-covered program may
not make an arrangement to get a discriminatory patent promise from a
third party for its customers, covering copies of the program (or
products that contain the program), if the arrangement requires the
distributor to make payment to the third party based on the extent of
its activity in conveying the program, and if the third party is itself
in the business of distributing software. Unlike the fourth paragraph,
which creates a legal defense for targets of patent aggression, the
consequence for violation of the fifth paragraph is termination of GPL
permissions for the distributor.
The business, technical, and patent cooperation agreement between
Microsoft and Novell announced in November 2006 has significantly
affected the development of Draft 3. The fourth and fifth paragraphs of
section 11 embody our response to the sort of threat represented by the
Microsoft/Novell deal, and are designed to protect users from such
deals, and prevent or deter the making of such deals.
The details of the agreements entered into between Microsoft and Novell,
though subject to eventual public disclosure through the securities
regulation system, have not been fully disclosed to this
point.\footnote{Lawyers employed by the Software Freedom Law Center,
which is counsel to the Free Software Foundation and other relevant free
software clients, were accorded limited access to the terms of the deal
under a non-disclosure agreement between SFLC and Novell. The reasons
for delay in the application of securities regulations requiring
publication of the relevant contracts are unrelated to the deal between
Microsoft and Novell.} It is a matter of public knowledge, however,
that the arrangement calls for Novell to pay a portion of the future
gross revenue of one of its divisions to Microsoft, and that (as one
other feature of a complex arrangement) Microsoft has promised Novell's
customers not to bring patent infringement actions against certain
specific copies of Novell's SUSE ``Linux''\footnote{This is a GNU/Linux
distribution, and is properly called SUSE GNU/Linux Enterprise Server.}
Enterprise Server product for which Novell receives revenue from the
user, so long as the user does not make or distribute additional copies
of SLES.
The basic harm that such an agreement can do is to make the free
software subject to it effectively proprietary. This result occurs to
the extent that users feel compelled, by the threat of the patent, to
get their copies in this way. So far, the Microsoft/Novell deal does
not seem to have had this result, or at least not very much: users do
not seem to be choosing Novell for this reason. But we cannot take for
granted that such threats will always fail to harm the community. We
take the threat seriously, and we have decided to act to block such
threats, and to reduce their potential to do harm. Such deals also
offer patent holders a crack through which to split the community.
Offering commercial users the chance to buy limited promises of patent
safety in effect invites each of them to make a separate peace with
patent aggressors, and abandon the rest of our community to its fate.
Microsoft has been restrained from patent aggression in the past by the
vocal opposition of its own enterprise customers, who now also use free
software systems to run critical applications. Public statements by
Microsoft concerning supposed imminent patent infringement actions have
spurred resistance from users Microsoft cannot afford to alienate. But
if Microsoft can gain royalties from commercial customers by assuring
them that \textit{their} copies of free software have patent licenses
through a deal between Microsoft and specific GNU/Linux vendors,
Microsoft would then be able to pressure each user individually, and
each distributor individually, to treat the software as proprietary. If
enough users succumb, it might eventually gain a position to terrify
noncommercial developers into abandoning the software entirely.
Preventing these harms is the goal of the new provisions of section 11.
The fourth paragraph deals with the most acute danger posed by
discrimination among customers, by ensuring that any party who
distributes others' GPL-covered programs, and makes promises of patent
safety limited to some but not all recipients of copies of those
specific programs, automatically extends its promises of patent safety
to cover all recipients of all copies of the covered works. This will
negate part of the harm of the Microsoft/Novell deal, for GPLv3-covered
software.
In addition to the present deal, however, GPLv3 must act to deter
similar future arrangements, and it cannot be assumed that all future
arrangements by Microsoft or other potential patent aggressors will
involve procuring the conveyance of the program by the party that grants
the discriminatory promises of patent safety. Therefore, we need the
fifth paragraph as well, which is aimed at parties that play the Novell
role in a different range of possible deals.
Drafting this paragraph was difficult because it is necessary to
distinguish between pernicious agreements and other kinds of agreements
which do not have an acutely harmful effect, such as patent
contributions, insurances, customary cross-license promises to
customers, promises incident to ordinary asset transfers, and standard
settlement practices. We believe that we have achieved this, but it is
hard to be sure, so we are considering making this paragraph apply only
to agreements signed in the future. If we do that, companies would only
need to structure future agreements in accord with the fifth paragraph,
and would not face problems from past agreements that cannot be changed
now. We are not yet convinced that this is necessary, and we plan to
ask for more comment on the question. This is why the date-based cutoff
is included in brackets.
One drawback of this cutoff date is that it would ``let Novell off''
from part of the response to its deal with Microsoft. However, this may
not be a great drawback, because the fourth paragraph will apply to that
deal. We believe it is sufficient to ensure either the deal's voluntary
modification by Microsoft or its reduction to comparative harmlessness.
Novell expected to gain commercial advantage from its patent deal with
Microsoft; the effects of the fourth paragraph in undoing the harm of
that deal will necessarily be visited upon Novell.
\section{GPLv3~\S12: Familiar as GPLv2 \S~7}
% FIXME: probably mostly still right, needs some updates, though.
The wording in the first sentence of section 12 has been revised
slightly to clarify that an agreement, such as a litigation settlement
agreement or a patent license agreement, is one of the ways in which
conditions may be ``imposed'' on a GPL licensee that may contradict the
conditions of the GPL, but which do not excuse the licensee from
compliance with those conditions. This change codifies what has been
our interpretation of GPLv2.
% FIXME: probably mostly still right, needs some updates, though.
We have removed the limited severability clause of GPLv2 section 7 as a
matter of tactical judgment, believing that this is the best way to ensure
that all provisions of the GPL will be upheld in court. We have also removed
the final sentence of GPLv2 section 7, which we consider to be unnecessary.
\section{GPLv3~\S13: The Great Affero Compromise}
The main purpose of clause 7b4 was to attain GPLv3 compatibility for the
additional condition of version 1 of the Affero GPL, with a view to
achieving compatibility for a future version, since version 1 was
incompatible with GPLv3.\footnote{Version 1 of the Affero GPL contains
its own copyleft clause, worded identically to that in GPLv2, which
conflicts with the copyleft clause in GPLv3. The Affero GPL permits
relicensing under versions of the GPL later than version 2, but only if
the later version ``includes terms and conditions substantially
equivalent to those of this license'' (Affero GPL, version 1, section
9). The Affero license was written with the expectation that its
additional requirement would be incorporated into the terms of GPLv3
itself, rather than being placeable on parts added to a covered work
through the mechanism of section 7 of GPLv3.} However, we wrote the
clause broadly enough to cover a range of other possible terms that
would differ from the Affero condition in their details. Draft 3 no
longer pursues the more ambitious goal of allowing compatibility for a
whole category of Affero-like terms. In place of 7b4, we have added a
new section 13 that simply permits GPLv3-covered code to be linked with
code covered by the forthcoming version 2 of the Affero GPL.
We have made this decision in the face of irreconcilable views from
different parts of our community. While we had known that many
commercial users of free software were opposed to the inclusion of a
mandatory Affero-like requirement in the body of GPLv3 itself, we were
surprised at their opposition to its availability through section 7.
Free software vendors allied to these users joined in their objections,
as did a number of free software developers arguing on ethical as well
as practical grounds.
Some of this hostility seemed to be based on a misapprehension that
Affero-like terms placed on part of a covered work would somehow extend
to the whole of the work.\footnote{It is possible that the presence of
the GPLv2-derived copyleft clause in the existing Affero GPL contributed
to this misunderstanding.} Our explanations to the contrary did little
to satisfy these critics; their objections to 7b4 instead evolved into a
broader indictment of the additional requirements scheme of section 7.
It was clear, however, that much of the concern about 7b4 stemmed from
its general formulation. Many were alarmed at the prospect of GPLv3
compatibility for numerous Affero-like licensing conditions,
unpredictable in their details but potentially having significant
commercial consequences.
On the other hand, many developers, otherwise sympathetic to the policy
goals of the Affero GPL, have objected to the form of the additional
requirement in that license. These developers were generally
disappointed with our decision to allow Affero-like terms through
section 7, rather than adopt a condition for GPLv3. Echoing their
concerns about the Affero GPL itself, they found fault with the wording
of the section 7 clause in both of the earlier drafts. We drafted 7b4
at a higher level than its Draft 1 counterpart based in part on comments
from these developers. They considered the Draft 1 clause too closely
tied to the Affero mechanism of preserving functioning facilities for
downloading source, which they found too restrictive of the right of
modification. The 7b4 rewording did not satisfy them, however. They
objected to its limitation to terms requiring compliance by network
transmission of source, and to the technically imprecise or inaccurate
use of the phrase ``same network session.''
We have concluded that any redrafting of the 7b4 clause would fail to
satisfy the concerns of both sets of its critics. The first group
maintains that GPLv3 should do nothing about the problem of public
use. The second group would prefer for GPLv3 itself to have an
Affero-like condition, but that seems to us too drastic. By permitting
GPLv3-covered code to be linked with code covered by version 2 of the
Affero GPL, the new section 13 honors our original commitment to
achieving GPL compatibility for the Affero license.
Version 2 of the Affero GPL is not yet published. We will work with
Affero, Inc., and with all other interested members of our community, to
complete the drafting of this license following the release of Draft 3,
with a goal of having a final version available by the time of our
adoption of the final version of GPLv3. We hope the new Affero license
will satisfy those developers who are concerned about the issue of
public use of unconveyed versions but who have concerns about the
narrowness of the condition in the existing Affero license.
As the second sentence in section 13 indicates, when a combined work is
made by linking GPLv3-covered code with Affero-covered code, the
copyleft on one part will not extend to the other part.\footnote{The
plan is that the additional requirement of the new Affero license will
state a reciprocal limitation.} That is to say, in such combinations,
the Affero requirement will apply only to the part that was brought into
the combination under the Affero license. Those who receive such a
combination and do not wish to use code under the Affero requirement may
remove the Affero-covered portion of the combination.
Those who criticize the permission to link with code under the Affero
GPL should recognize that most other free software licenses also permit
such linking.
\section{GPLv3~\S14: So, When's GPLv4?}
\label{GPLv3s14}
% FIXME Say more
No substantive change has been made in section 14. The wording of the section
has been revised slightly to make it clearer.
% FIXME; proxy
\section{GPLv3~\S15--17: Warranty Disclaimers and Liability Limitation}
No substantive changes have been made in sections 15 and 16.
% FIXME: more, plus 17
% FIXME: Section header needed here about choice of law.
% FIXME: reword into tutorial
Some have asked us to address the difficulties of internationalization
by including, or permitting the inclusion of, a choice of law
provision. We maintain that this is the wrong approach. Free
software licenses should not contain choice of law clauses, for both
legal and pragmatic reasons. Choice of law clauses are creatures of
contract, but the substantive rights granted by the GPL are defined
under applicable local copyright law. Contractual free software
licenses can operate only to diminish these rights. Choice of law
clauses also raise complex questions of interpretation when works of
software are created by combination and extension. There is also the
real danger that a choice of law clause will specify a jurisdiction
that is hostile to free software principles.
% FIXME: reword into tutorial, \ref to section 7.
Our revised version of section 7 makes explicit our view that the
inclusion of a choice of law clause by a licensee is the imposition of
an additional requirement in violation of the GPL. Moreover, if a
program author or copyright holder purports to supplement the GPL with
a choice of law clause, section 7 now permits any licensee to remove
that clause.
% FIXME: does this need to be a section, describing how it was out then in
% then out then in? :)
We have removed from this draft the appended section on ``How to Apply These
Terms to Your New Programs.'' For brevity, the license document can instead
refer to a web page containing these instructions as a separate document.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\chapter{The Lesser GPL}
As we have seen in our consideration of the GPL, its text is specifically
designed to cover all possible derivative works under copyright law. Our
goal in designing GPL was to make sure that any derivative work of GPL'd
software was itself released under GPL when distributed. Reaching as far
as copyright law will allow is the most direct way to reach that goal.
However, while the strategic goal is to bring as much Free Software
into the world as possible, particular tactical considerations
regarding software freedom dictate different means. Extending the
copyleft effect as far as copyright law allows is not always the most
prudent course in reaching the goal. In particular situations, even
those of us with the goal of building a world where all published
software is Free Software realize that full copyleft does not best
serve us. The GNU Lesser General Public License (``GNU LGPL'') was
designed as a solution for such situations.
\section{The First LGPL'd Program}
The first example that FSF encountered where such altered tactics were
needed was when work began on the GNU C Library. The GNU C Library would
become (and today, now is) a drop-in replacement for existing C libraries.
On a Unix-like operating system, C is the lingua franca and the C library
is an essential component for all programs. It is extremely difficult to
construct a program that will run with ease on a Unix-like operating
system without making use of services provided by the C library --- even
if the program is written in a language other than C\@. Effectively, all
user application programs that run on any modern Unix-like system must
make use of the C library.
By the time work began on the GNU implementation of the C libraries, there
were already many C libraries in existence from a variety of vendors.
Every proprietary Unix vendor had one, and many third parties produced
smaller versions for special purpose use. However, our goal was to create
a C library that would provide equivalent functionality to these other C
libraries on a Free Software operating system (which in fact happens today
on modern GNU/Linux systems, which all use the GNU C Library).
Unlike existing GNU application software, however, the licensing
implications of releasing the GNU C Library (``glibc'') under GPL were
somewhat different. Applications released under GPL would never
themselves become part of proprietary software. However, if glibc were
released under GPL, it would require that any application distributed for
the GNU/Linux platform be released under GPL\@.
Since all applications on a Unix-like system depend on the C library, it
means that they must link with that library to function on the system. In
other words, all applications running on a Unix-like system must be
combined with the C library to form a new whole derivative work that is
composed of the original application and the C library. Thus, if glibc
were GPL'd, each and every application distributed for use on GNU/Linux
would also need to be GPL'd, since to even function, such applications
would need to be combined into larger derivative works by linking with
glibc.
At first glance, such an outcome seems like a windfall for Free Software
advocates, since it stops all proprietary software development on
GNU/Linux systems. However, the outcome is a bit more subtle. In a world
where many C libraries already exist, many of which could easily be ported
to GNU/Linux, a GPL'd glibc would be unlikely to succeed. Proprietary
vendors would see the excellent opportunity to license their C libraries
to anyone who wished to write proprietary software for GNU/Linux systems.
The de-facto standard for the C library on GNU/Linux would likely be not
glibc, but the most popular proprietary one.
Meanwhile, the actual goal of releasing glibc under GPL --- to ensure no
proprietary applications on GNU/Linux --- would be unattainable in this
scenario. Furthermore, users of those proprietary applications would also
be users of a proprietary C library, not the Free glibc.
The Lesser GPL was initially conceived to handle this scenario. It was
clear that the existence of proprietary applications for GNU/Linux was
inevitable. Since there were so many C libraries already in existence, a
new one under GPL would not stop that tide. However, if the new C library
were released under a license that permitted proprietary applications
to link with it, but made sure that the library itself remained Free,
an ancillary goal could be met. Users of proprietary applications, while
they would not have the freedom to copy, share, modify and redistribute
the application itself, would have the freedom to do so with respect to
the C library.
There was no way the license of glibc could stop or even slow the creation
of proprietary applications on GNU/Linux. However, loosening the
restrictions on the licensing of glibc ensured that nearly all proprietary
applications at least used a Free C library rather than a proprietary one.
This trade-off is central to the reasoning behind the LGPL\@.
Of course, many people who use the LGPL today are not thinking in these
terms. In fact, they are often choosing the LGPL because they are looking
for a ``compromise'' between the GPL and the X11-style liberal licensing.
However, understanding FSF's reasoning behind the creation of the LGPL is
helpful when studying the license.
\section{What's the Same?}
Much of the text of the LGPL is identical to the GPL\@. As we begin our
discussion of the LGPL, we will first eliminate the sections that are
identical, or that have the minor modification changing the word
``Program'' to ``Library.''
First, LGPLv2.1~\S1, the rules for verbatim copying of source, are
equivalent to those in GPLv2~\S1.
Second, LGPLv2.1~\S8 is equivalent GPLv2~\S4\@. In both licenses, this
section handles termination in precisely the same manner.
LGPLv2.1~\S9 is equivalent to GPLv2~\S5\@. Both sections assert that
the license is a copyright license, and handle the acceptance of those
copyright terms.
LGPLv2.1~\S10 is equivalent to GPLv2~\S6. They both protect the
distribution system of Free Software under these licenses, to ensure that
up, down, and throughout the distribution chain, each recipient of the
software receives identical rights under the license and no other
restrictions are imposed.
LGPLv2.1~\S11 is GPLv2~\S7. As discussed, it is used to ensure that
other claims and legal realities, such as patent licenses and court
judgments, do not trump the rights and permissions granted by these
licenses, and requires that distribution be halted if such a trump is
known to exist.
LGPLv2.1~\S12 adds the same features as GPLv2~\S8. These sections are
used to allow original copyright holders to forbid distribution in
countries with draconian laws that would otherwise contradict these
licenses.
LGPLv2.1~\S13 sets up FSF as the steward of the LGPL, just as GPLv2~\S9
does for GPL. Meanwhile, LGPLv2.1~\S14 reminds licensees that copyright
holders can grant exceptions to the terms of LGPL, just as GPLv2~\S10
reminds licensees of the same thing.
Finally, the assertions of no warranty and limitations of liability are
identical; thus LGPLv2.1~\S15 and LGPLv2.1~\S16 are the same as GPLv2~\S11 and \S
12.
As we see, the entire latter half of the license is identical.
The parts which set up the legal boundaries and meta-rules for the license
are the same. It is our intent that the two licenses operate under the
same legal mechanisms and are enforced precisely the same way.
We strike a difference only in the early portions of the license.
Namely, in the LGPL we go into deeper detail of granting various permissions to
create derivative works, so the redistributors can make
some proprietary derivatives. Since we simply do not allow the
license to stretch as far as copyright law does regarding what
derivative works must be relicensed under the same terms, we must go
further to explain which derivative works we will allow to be
proprietary. Thus, we'll see that the front matter of the LGPL is a
bit more wordy and detailed with regards to the permissions granted to
those who modify or redistribute the software.
\section{Additions to the Preamble}
Most of LGPL's Preamble is identical, but the last seven paragraphs
introduce the concepts and reasoning behind creation of the license,
presenting a more generalized and briefer version of the story with which
we began our consideration of LGPL\@.
In short, FSF designed LGPL for those edge cases where the freedom of the
public can better be served by a more lax licensing system. FSF doesn't
encourage use of LGPL automatically for any software that happens to be a
library; rather, FSF suggests that it only be used in specific cases, such
as the following:
\begin{itemize}
\item To encourage the widest possible use of a Free Software library, so
it becomes a de-facto standard over similar, although not
interface-identical, proprietary alternatives
\item To encourage use of a Free Software library that already has
interface-identical proprietary competitors that are more developed
\item To allow a greater number of users to get freedom, by encouraging
proprietary companies to pick a Free alternative for its otherwise
proprietary products
\end{itemize}
LGPL's preamble sets forth the limits to which the license seeks to go in
chasing these goals. LGPL is designed to ensure that users who happen to
acquire software linked with such libraries have full freedoms with
respect to that library. They should have the ability to upgrade to a newer
or modified Free version or to make their own modifications, even if they
cannot modify the primary software program that links to that library.
Finally, the preamble introduces two terms used throughout the license to
clarify between the different types of derivative works: ``works that use
the library,'' and ``works based on the library.'' Unlike GPL, LGPL must
draw some lines regarding derivative works. We do this here in this
license because we specifically seek to liberalize the rights afforded to
those who make derivative works. In GPL, we reach as far as copyright law
allows. In LGPL, we want to draw a line that allows some derivative works
copyright law would otherwise prohibit if the copyright holder exercised
his full permitted controls over the work.
\section{An Application: A Work that Uses the Library}
In the effort to allow certain proprietary derivative works and prohibit
others, LGPL distinguishes between two classes of derivative works:
``works based on the library,'' and ``works that use the library.'' The
distinction is drawn on the bright line of binary (or runtime) derivative
works and source code derivatives. We will first consider the definition
of a ``work that uses the library,'' which is set forth in LGPLv2.1~\S5.
We noted in our discussion of GPLv2~\S3 (discussed in
Section~\ref{GPLv2s3} of this document) that binary programs when
compiled and linked with GPL'd software are derivative works of that GPL'd
software. This includes both linking that happens at compile-time (when
the binary is created) or at runtime (when the binary -- including library
and main program both -- is loaded into memory by the user). In GPL,
binary derivative works are controlled by the terms of the license (in GPLv2~\S3),
and distributors of such binary derivatives must release full
corresponding source\@.
In the case of LGPL, these are precisely the types of derivative works
we wish to permit. This scenario, defined in LGPL as ``a work that uses
the library,'' works as follows:
\newcommand{\workl}{$\mathcal{L}$}
\newcommand{\lplusi}{$\mathcal{L\!\!+\!\!I}$}
\begin{itemize}
\item A new copyright holder creates a separate and independent work,
\worki{}, that makes interface calls (e.g., function calls) to the
LGPL'd work, called \workl{}, whose copyright is held by some other
party. Note that since \worki{} and \workl{} are separate and
independent works, there is no copyright obligation on this new copyright
holder with regard to the licensing of \worki{}, at least with regard to
the source code.
\item The new copyright holder, for her software to be useful, realizes
that it cannot run without combining \worki{} and \workl{}.
Specifically, when she creates a running binary program, that running
binary must be a derivative work, called \lplusi{}, that the user can
run.
\item Since \lplusi{} is a derivative work of both \worki{} and \workl{},
the license of \workl{} (the LGPL) can put restrictions on the license
of \lplusi{}. In fact, this is what LGPL does.
\end{itemize}
We will talk about the specific restrictions LGPLv2.1 places on ``works
that use the library'' in detail in Section~\ref{lgpl-section-6}. For
now, focus on the logic related to how the LGPLv2.1 places requirements on
the license of \lplusi{}. Note, first of all, the similarity between
this explanation and that in Section~\ref{separate-and-independent},
which discussed the combination of otherwise separate and independent
works with GPL'd code. Effectively, what LGPLv2.1 does is say that when a
new work is otherwise separate and independent, but has interface
calls out to an LGPL'd library, then it is considered a ``work that
uses the library.''
In addition, the only reason that LGPLv2.1 has any control over the licensing
of a ``work that uses the library'' is for the same reason that GPL has
some say over separate and independent works. Namely, such controls exist
because the {\em binary combination\/} (\lplusi{}) that must be created to
make the separate work (\worki{}) at all useful is a derivative work of
the LGPLv2.1'd software (\workl{}).
Thus, a two-question test that will help indicate if a particular work is
a ``work that uses the library'' under LGPLv2.1 is as follows:
\begin{enumerate}
\item Is the source code of the new copyrighted work, \worki{}, a
completely independent work that stands by itself, and includes no
source code from \workl{}?
\item When the source code is compiled, does it create a derivative work
by combining with \workl{}, either by static (compile-time) or dynamic
(runtime) linking, to create a new binary work, \lplusi{}?
\end{enumerate}
If the answers to both questions are ``yes,'' then \worki{} is most likely
a ``work that uses the library.'' If the answer to the first question
``yes,'' but the answer to the second question is ``no,'' then most likely
\worki{} is neither a ``work that uses the library'' nor a ``work based on
the library.'' If the answer to the first question is ``no,'' but the
answer to the second question is ``yes,'' then an investigation into
whether or not \worki{} is in fact a ``work based on the library'' is
warranted.
\section{The Library, and Works Based On It}
In short, a ``work based on the library'' could be defined as any
derivative work of LGPL'd software that cannot otherwise fit the
definition of a ``work that uses the library.'' A ``work based on the
library'' extends the full width and depth of copyright derivative works,
in the same sense that GPL does.
Most typically, one creates a ``work based on the library'' by directly
modifying the source of the library. Such a work could also be created by
tightly integrating new software with the library. The lines are no doubt
fuzzy, just as they are with GPL'd works, since copyright law gives us no
litmus test for derivative works of a software program.
Thus, the test to use when considering whether something is a ``work
based on the library'' is as follows:
\begin{enumerate}
\item Is the new work, when in source form, a derivative work under
copyright law of the LGPL'd work?
\item Is there no way in which the new work fits the definition of a
``work that uses the library''?
\end{enumerate}
If the answer is ``yes'' to both these questions, then you most likely
have a ``work based on the library.'' If the answer is ``no'' to the
first but ``yes'' to the second, you are in a gray area between ``work
based on the library'' and a ``work that uses the library.''
In our years of work with the LGPLv2.1, however, we have never seen a work
of software that was not clearly one or the other; the line is quite
bright. At times, though, we have seen cases where a derivative work
appeared in some ways to be a work that used the library and in other
ways a work based on the library. We overcame this problem by
dividing the work into smaller subunits. It was soon discovered that
what we actually had were three distinct components: the original
LGPL'd work, a specific set of works that used that library, and a
specific set of works that were based on the library. Once such
distinctions are established, the licensing for each component can be
considered independently and the LGPLv2.1 applied to each work as
prescribed.
\section{Subtleties in Defining the Application}
In our discussion of the definition of ``works that use the library,'' we
left out a few more complex details that relate to lower-level programming
details. The fourth paragraph of LGPLv2.1~\S5 covers these complexities,
and it has been a source of great confusion. Part of the confusion comes
because a deep understanding of how compiler programs work is nearly
mandatory to grasp the subtle nature of what LGPLv2.1~\S5, \P 4 seeks to
cover. It helps some to note that this is a border case that we cover in
the license only so that when such a border case is hit, the implications
of using LGPL continue in the expected way.
To understand this subtle point, we must recall the way that a compiler
operates. The compiler first generates object code, which are the binary
representations of various programming modules. Each of those modules is
usually not useful by itself; it becomes useful to a user of a full program
when those modules are {\em linked\/} into a full binary executable.
As we have discussed, the assembly of modules can happen at compile-time
or at runtime. Legally, there is no distinction between the two --- both
create a derivative work by copying and combining portions of one work and
mixing them with another. However, under LGPL, there is a case in the
compilation process where the legal implications are different.
Specifically, while we know that a ``work that uses the library'' is one
whose final binary is a derivative work, but whose source is not, there
are cases where the object code --- that intermediate step between source
and final binary --- is a derivative work created by copying verbatim code
from the LGPL'd software.
For efficiency, when a compiler turns source code into object code, it
sometimes places literal portions of the copyrighted library code into the
object code for an otherwise separate independent work. In the normal
scenario, the derivative would not be created until final assembly and
linking of the executable occurred. However, when the compiler does this
efficiency optimization, at the intermediate object code step, a
derivative work is created.
LGPLv2.1~\S5\P4 is designed to handle this specific case. The intent of
the license is clearly that simply compiling software to ``make use'' of
the library does not in itself cause the compiled work to be a ``work
based on the library.'' However, since the compiler copies verbatim,
copyrighted portions of the library into the object code for the otherwise
separate and independent work, it would actually cause that object file to be a
``work based on the library.'' It is not FSF's intent that a mere
compilation idiosyncrasy would change the requirements on the users of the
LGPLv2.1'd software. This paragraph removes that restriction, allowing the
implications of the license to be the same regardless of the specific
mechanisms the compiler uses underneath to create the ``work that uses the
library.''
As it turns out, we have only once had anyone worry about this specific
idiosyncrasy, because that particular vendor wanted to ship object code
(rather than final binaries) to their customers and was worried about
this edge condition. The intent of clarifying this edge condition is
primarily to quell the worries of software engineers who understand the
level of verbatim code copying that a compiler often does, and to help
them understand that the full implications of LGPLv2.1 are the same regardless
of the details of the compilation progress.
\section{LGPLv2.1~\S6 \& LGPLv2.1~\S5: Combining the Works}
\label{lgpl-section-6}
Now that we have established a good working definition of works that
``use'' and works that ``are based on'' the library, we will consider the
rules for distributing these two different works.
The rules for distributing ``works that use the library'' are covered in
LGPLv2.1~\S6\@. LGPLv2.1~\S6 is much like GPLv2~\S3, as it requires the release
of source when a binary version of the LGPL'd software is released. Of
course, it only requires that source code for the library itself be made
available. The work that ``uses'' the library need not be provided in
source form. However, there are also conditions in LGPLv2.1~\S6 to make sure
that a user who wishes to modify or update the library can do so.
LGPLv2.1~\S6 lists five choices with regard to supplying library source
and granting the freedom to modify that library source to users. We
will first consider the option given by \S~6(b), which describes the
most common way currently used for LGPLv2.1 compliance on a ``work that
uses the library.''
LGPLv2.1~\S6(b) allows the distributor of a ``work that uses the library'' to
simply use a dynamically linked, shared library mechanism to link with the
library. This is by far the easiest and most straightforward option for
distribution. In this case, the executable of the work that uses the
library will contain only the ``stub code'' that is put in place by the
shared library mechanism, and at runtime the executable will combine with
the shared version of the library already resident on the user's computer.
If such a mechanism is used, it must allow the user to upgrade and
replace the library with interface-compatible versions and still be able
to use the ``work that uses the library.'' However, all modern shared
library mechanisms function as such, and thus LGPLv2.1~\S6(b) is the simplest
option, since it does not even require that the distributor of the ``work
2based on the library'' ship copies of the library itself.
LGPLv2.1~\S6(a) is the option to use when, for some reason, a shared library
mechanism cannot be used. It requires that the source for the library be
included, in the typical GPL fashion, but it also has a requirement beyond
that. The user must be able to exercise her freedom to modify the library
to its fullest extent, and that means recombining it with the ``work based
on the library.'' If the full binary is linked without a shared library
mechanism, the user must have available the object code for the ``work
based on the library,'' so that the user can relink the application and
build a new binary.
The remaining options in LGPLv2.1~\S6 are very similar to the other choices
provided by GPLv2~\S3. There are some additional options, but time does
not permit us in this course to go into those additional options. In
almost all cases of distribution under LGPL, either LGPLv2.1~\S6(a) or LGPLv2.1~\S6(b) are
exercised.
\section{Distribution of the Combined Works}
Essentially, ``works based on the library'' must be distributed under the
same conditions as works under full GPL\@. In fact, we note that
LGPLv2.1~\S2 is nearly identical in its terms and requirements to GPLv2~\S2.
There are again subtle differences and additions, which time does not
permit us to cover in this course.
\section{And the Rest}
The remaining variations between LGPL and GPL cover the following
conditions:
\begin{itemize}
\item Allowing a licensing ``upgrade'' from LGPL to GPL\@ (in LGPLv2.1~\S3)
\item Binary distribution of the library only, covered in LGPLv2.1~\S4,
which is effectively equivalent to LGPLv2.1~\S3
\item Creating aggregates of libraries that are not derivative works of
each other, and distributing them as a unit (in LGPLv2.1~\S7)
\end{itemize}
Due to time constraints, we cannot cover these additional terms in detail,
but they are mostly straightforward. The key to understanding LGPLv2.1 is
understanding the difference between a ``work based on the library'' and a
``work that uses the library.'' Once that distinction is clear, the
remainder of LGPLv2.1 is close enough to GPL that the concepts discussed in
our more extensive GPL unit can be directly applied.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\chapter{Integrating the GPL into Business Practices}
Since GPL'd software is now extremely prevalent through the industry, it
is useful to have some basic knowledge about using GPL'd software in
business and how to build business models around GPL'd software.
\section{Using GPL'd Software In-House}
As discussed in Sections~\ref{GPLv2s0} and~\ref{GPLv2s5} of this tutorial,
the GPL only governs the activities of copying, modifying and
distributing software programs that are not governed by the license.
Thus, in FSF's view, simply installing the software on a machine and
using it is not controlled or limited in any way by GPL\@. Using Free
Software in general requires substantially fewer agreements and less
license compliance activity than any known proprietary software.
Even if a company engages heavily in copying the software throughout the
enterprise, such copying is not only permitted by GPLv2~\S\S1 and 3, but it is
encouraged! If the company simply deploys unmodified (or even modified)
Free Software throughout the organization for its employees to use, the
obligations under the license are very minimal. Using Free Software has a
substantially lower cost of ownership --- both in licensing fees and in
licensing checking and handling -- than the proprietary software
equivalents.
\section{Business Models}
\label{Business Models}
Using Free Software in house is certainly helpful, but a thriving
market for Free Software-oriented business models also exists. There is the
traditional model of selling copies of Free Software distributions.
Many companies, including IBM and Red Hat, make substantial revenue
from this model. IBM primarily chooses this model because they have
found that for higher-end hardware, the cost of the profit made from
proprietary software licensing fees is negligible. The real profit is
in the hardware, but it is essential that software be stable, reliable
and dependable, and the users be allowed to have unfettered access to
it. Free Software, and GPL'd software in particular (because IBM can
be assured that proprietary versions of the same software will not
exists to compete on their hardware) is the right choice.
Red Hat has actually found that a ``convenience fee'' for Free Software,
when set at a reasonable price (around \$60 or so), can produce some
profit. Even though Red Hat's system is fully downloadable on their
Web site, people still go to local computer stores and buy copies of their
box set, which is simply a printed version of the manual (available under
a Free license as well) and the Free Software system it documents.
\medskip
However, custom support, service, and software improvement contracts
are the most widely used models for GPL'd software. The GPL is
central to their success, because it ensures that the code base
remains common, and that large and small companies are on equal
footing for access to the technology. Consider, for example, the GNU
Compiler Collection (GCC). Cygnus Solutions, a company started in the
early 1990s, was able to grow steadily simply by providing services
for GCC --- mostly consisting of new ports of GCC to different or new,
embedded targets. Eventually, Cygnus was so successful that
it was purchased by Red Hat where it remains a profitable division.
However, there are very small companies like CodeSourcery, as well as
other medium-sized companies like MontaVista and OpenTV that compete in
this space. Because the code-base is protect by GPL, it creates and
demands industry trust. Companies can cooperate on the software and
improve it for everyone. Meanwhile, companies who rely on GCC for their
work are happy to pay for improvements, and for ports to new target
platforms. Nearly all the changes fold back into the standard
versions, and those forks that exist remain freely available.
\medskip
\label{Proprietary Relicensing}
A final common business model that is perhaps the most controversial is
proprietary relicensing of a GPL'd code base. This is only an option for
software in which a particular entity is the sole copyright holder. As
discussed earlier in this tutorial, a copyright holder is permitted under
copyright law to license a software system under her copyright as many
different ways as she likes to as many different parties as she wishes.
Some companies, such as MySQL AB and TrollTech, use this to their
financial advantage with regard to a GPL'd code base. The standard
version is available from the company under the terms of the GPL\@.
However, parties can purchase separate proprietary software licensing for
a fee.
This business model is problematic because it means that the GPL'd code
base must be developed in a somewhat monolithic way, because volunteer
Free Software developers may be reluctant to assign their copyrights to
the company because it will not promise to always and forever license the
software as Free Software. Indeed, the company will surely use such code
contributions in proprietary versions licensed for fees.
\section{Ongoing Compliance}
GPL compliance is in fact a very simple matter -- much simpler than
typical proprietary software agreements and EULAs. Usually, the most
difficult hurdle is changing from a proprietary software mindset to one
that seeks to foster a community of sharing and mutual support. Certainly
complying with the GPL from a users' perspective gives substantially fewer
headaches than proprietary license compliance.
For those who go into the business of distributing {\em modified\\}
versions of GPL'd software, the burden is a bit higher, but not by
much. The glib answer is that by releasing the whole product as Free
Software, it is always easy to comply with the GPL. However,
admittedly to the dismay of FSF, many modern and complex software
systems are built using both proprietary and GPL'd components that are
not legally derivative works of each other. Sometimes, it is easier simply to
improve existing GPL'd application than to start from scratch. In
exchange for that benefit, the license requires that the modifier give
back to the commons that made the work easier in the first place. It is a
reasonable trade-off and a way to help build a better world while also
making a profit.
Note that FSF does provide services to assist companies who need
assistance in complying with the GPL. You can contact FSF's GPL
Compliance Labs at $<$compliance@fsf.org$>$.
If you are particularly interested in matters of GPL compliance, we
recommend the second course in this series, {\em GPL Compliance Case
Studies and Legal Ethics in Free Software Licensing\/}, in which we
discuss some real GPL violation cases that FSF has worked to resolve.
Consideration of such cases can help give insight on how to handle GPL
compliance in new situations.
% =====================================================================
% END OF FIRST DAY SEMINAR SECTION
% =====================================================================