56250
Comment: EBNF example table: Fix LGPL row
|
57703
Add to EBNF new topic ''Notes:''
|
Deletions are marked like this. | Additions are marked like this. |
Line 385: | Line 385: |
Notes: The extra plus-sign in {{{<License version>["+"]}}} means that the License contains clauses similar to GPL's "version N or (at your option) any later version". |
|
Line 402: | Line 398: |
Notes: The extra plus-sign in {{{<License version>["+"]}}} means that the License contains clauses similar to GPL's "version N or (at your option) any later version". The GPL/LGPL {{{BY-OpenSSL}}} clarification gives permission in cases where GPL code is linked with OpenSSL library. For more information, see page [http://www.gnome.org/~markmc/openssl-and-the-gpl.html "The OpenSSL License and The GPL"] by Mark McLoughlin and message [http://lists.debian.org/debian-legal/2004/05/msg00595.html "middleman software license conflicts with OpenSSL"] by Mark McLoughlin in Debian legal mainling list. The License needs to include following or close to similar excerpt to qualify for {{{BY-OpenSSL}}} which allows linking with OpenSSL: {{{ In addition, as a special exception, the copyright holders give permission to link the code of portions of this program with the OpenSSL library under certain conditions as described in each individual source file, and distribute linked combinations including the two. You must obey the GNU General Public License in all respects for all of the code used other than OpenSSL. If you modify file(s) with this exception, you may extend this exception to your version of the file(s), but you are not obligated to do so. If you do not wish to do so, delete this exception statement from your version. If you delete this exception statement from all source files in the program, then also delete it here. }}} |
|
Line 664: | Line 683: |
* 2008-11-21: Add to EBNF new topic ''Notes:''. Add OpenSSL exception excerpt and cite Mark McLoughlin. |
Comment: I have been considering that we may need to alter how we develop this standard. Editing on a wiki, with inline comments is pretty frustrating. It becomes troublesome when people make large changes without much discussion beforehand. What do people think to moving or locking this place so that it cannot be arbitrarily edited and starting a mailing list for discussion, so that we can reach consensus about changes before making them. I would like to have an threadable and searchable discussion archive for this. -- ?NoahSlater
Comment: I think that the proposal got developed and popularised enough that a [http://dep.debian.net/deps/dep0/ DEP] could be proposed (but after Lenny release, please!). The discussion of the DEP will be more efficient to resolve out the remaining issues that are debated now using the wiki as a medium.-- CharlesPlessy
Comment: This sounds like a fine idea. I think that the wiki is starting to hinder, rather than promote, the development of this standard. Let's wait for Lenny and move forward from there. -- ?NoahSlater
This is a proposal to make debian/copyright machine-interpretable. This file is one of the most important files in Debian packaging, yet its existing format is vague and varies tremendously across packages, making it difficult to automatically parse.
This is not a proposal to change the policy in the short term.
?TableOfContents
Rationale
The diversity of free software licenses means that Debian needs to care not only about the freeness of a given work, but also its license's compatibility with the other parts of Debian it uses.
The arrival of the GPL version 3, its incompatibility with version 2, and our inability to spot the software where the incompatibility might be problematic is one prominent occurrence of this limitation.
There are earlier precedents, also. One is the GPL/OpenSSL incompatibility. Apart from grepping debian/copyright, which is prone to numerous false positives (packaging under the GPL but software under another license) or negatives (GPL software but with an "OpenSSL special exception" dual licensing form), there is no reliable way to know which software in Debian might be problematic.
And there is more to come. There are issues with shipping GPLv2-only software with a CDDL operating system such as Nexenta. The GPL version 3 solves this issue, but not all GPL software can switch to it and we have no way to know how much of Debian should be stripped from such a system.
Comment: apparently, [http://fedoraproject.org/wiki/Licensing Fedora started a very similar project]: on of my upstreams (the Samba Team) pointed me to the very few differences in the way to name licenses, particularly the short-form and the method to combine licenses. —ChristianPerrier
Compatibility and Human-Readability
The file must be encoded as UTF-8 and strictly formatted as a superset of RFC2822 including significant newlines. Free-form text is not allowed.
The debian/copyright file must be machine-interpretable, yet human-readable, while communicating all mandated upstream information, copyright notices and licensing details.
For the sake of human-readability this proposal avoids any complex field names or syntax rules.
Lintian
You can discuss implementation details in [http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=478930 bug 478930] -- ?MathieuParent
Implementation
Sections
Header Section (Once)
The header should be rfc2822 compliant, consisting of multiple fields.
- format specification
- only once
- mandatory
Suggested name: Format-Specification
- Suggested format: URI of the format specification, such as:
http://wiki.debian.org/Proposals/CopyrightFormat?action=recall&rev=REVISION
- Note that the unwieldy length of the URL should be solved in future by hosting the specification at a shorter URL (including specification version).
Currently (2008-07-25) this field can trigger a lintian warning about excessive line length; an exception for the Format-Specification line is pending in [http://bugs.debian.org/491302 bug 491302].
- name of the software as spelled upstream
- only once
- optional
Suggested name: Upstream-Name
- Suggested format: single line (in most cases a single word)
- preferred way(s) to reach current upstream maintainer
- only once
- optional
Suggested name: Upstream-Maintainer
- Suggested format: line(s) of RFC2822 address or URIs or free text, one line per preferred way
- where the upstream sources (if any) were downloaded from
- only once
- optional
Suggested name: Upstream-Source
- Suggested format: line(s) of URIs, one line per upstream source
- where the upstream VCS repository can be browsed online
- only once
- optional
Suggested name: Upstream-Vcs-Browser
- Suggested format: line(s) of HTTP URLs, one line per upstream Version Control Repository
- where the upstream VCS repository can be accessed
- only once
- optional
Suggested name: Upstream-Vcs-URI
- Suggested format: line(s) of MIME type definition. Example:
Upstream-Vcs-URI: type=git; uri=http://git.example.com/project.git
Comment: I am not convinced that Upstream-Vcs-Browser or Upstream-Vcs-URI belong in debian/copyright. Perhaps debian/READEME.source would be a better place for this. In either case, this is way too complex as it stands. What purpose does it serve? Can we just stick with vanila URIs please? -- Noah Slater
Comment: It is machine readable data, similar to Upstream-Source. MIME-like tags are easy for machine processing. -- JariAalto
Comment: Jari, the question (as I understand it) is not about the *format* of this proposed data, but rather whether it belongs in debian/copyright at all, rather than, say, debian/README.source. I'm in agreement with Noah on this: it doesn't seem debian/copyright is the right place for this. —BenFinney
Comment: Correct. I was (perhaps confusingly) picking up on two issues here. The proposal for Upstream-Vcs-URI as it stands is way too complex for any use case I can imagine. Who would care about the "priority=50" value, for example? Being machine readable alone is not justification for it's inclusion in any part of the Debian packaging. The second issue I see is that I think this belongs in debian/README.source. -- ?NoahSlater
Comment: Well, the debian/copyright file is shown in the packages.debian.org, whereas the content of debian/README.source is not. I would see the Upstream-Source field being analogous to Upstream-Vcs-URI only differing in scheme of more modern access method. The upstream VCS information and download location is tightly coupled and it would be beneficial to keep them at the same place. The fields were only examples of how the information is extendable if presented using the MIME notation (most likely there only will be need for type and uri). -- JariAalto
Comment: I fail to see how the upstream version control system is at all relevant to the copyright file at all. It makes no odds what is and isn't displayed on any given website. The purpose of this file is a legal one, not to provide every single possible way of access the source, that information belongs, very specifically, in debian/README.source. I fail to see who this would be beneficial to. Again, using "MIME notation" is extremely over the top for what we are trying to achieve here.
Comment: I'm not sure where there was any requirement for "every single source". If the Copyright is to list location of the sources, the field can be used for cases where sources are only available through VCS. -- JariAalto
Comment: And of course, one of the best things about URIs is that they can be used to point to almost anything, from a webpage, a person, unicorn or the moon. If package wants to point to a VCS using the Upstream-Source they can do, by using whatever URI they feel like. We should not be adding additional fields, just for every different type of source repository. HTTP, FTP, WebDAV, Subversion, Git, or whatever. -- ?NoahSlater
Comment: There was no poposal to "additional fields...for every different type". One field can cover them all. That's the beaty of the standard MIME. Clearly defined field makes it possible to do automated link validity checks, tarball creations etc. -- JariAalto
Comment: The location of the upstream VCS is orthogonal to copyright issues because packages aren't (usually) packaged out of VCS but from a tarball. Moreover, VCS data is already in debian/control so including it in debian/copyright is duplicating information and redundancy == bad. All that is needed for a statement of copyright information is the Upstream-Source field. Additionally, this MIME-based specification would look like it's truly over-engineered. How many packages actually have more than one VCS that is being used for their data at the same time? How many of them are packaged out of the VCS and not from tarballs that are released? And if their source is strangely split and difficult to get hold of, then we already have README.Source to describe this. We should stick to the KISS principle with all this; complexity for complexity's sake will only hinder the adoption (and usefulness) of this file format. My feeling is that this is the wrong format for the information and that it is also the wrong place to record it. -- StuartPrescott
Comment: "....VCS data is already in debian/control so including it in debian/copyright is duplicating information and redundancy". Correction: the debian/control Vcs-* headers records Debian packaging repository, not the Upstream's one discussed here. There is no duplication. -- JariAalto
Comment: My comments were apropos Upstream-Vcs-Browser and Upstream-Vcs-URI, which are two more fields than was originally in this proposal. Like has been mentioned before, these additional fields, and the MIME information is completely irrelevant to the purpose of this file and should be dropped. My question would be, if you add "category=vcs; type=git;", what or who is going to use this information? And for any reply, the follow up question would be, should that agent be looking at this file for that information? -- ?NoahSlater
Comment: "two more fields" is hardly a problem; and theses are even optional. MIME is standard methodology in the context of the referred underlying layout RFC2822 where the ?CopyrightFormat proposal stems from. "who is going to use this information" is answered by the tools that make use of the machine readable copyright file. I can think of the same tools that uses the Upstream-Source. Field Upstream-Vcs-URI is fundamentally no different; it just happens to point to VCS instead of tarball. In Debian there exists snapshots packaged directly from VCS. We must also keep in mind the advancements of distributed VCS (Hg, Bzr, Git) that will surely gain gound more and more. Sources can probably in near future be packaged directly from VCS by following the attached tags (see Alioth Git repositories for examples) -- JariAalto
Comment: Yes, two more fields is very definitely a problem if they are not required. Any kind of additional complexity is a problem, there is no threshold of value below which unneeded features can be freely added to this specification. Also, you're missing the point of my question about who is going to use this information. I am not arguing that the information isn't broadly useful, it clearly is. I am arguing that it is not useful within the context of this file, which is only about copyright information. A single field for specifying the upstream source location is all that is needed. Anything more is surplus to requirements and should be dropped. -- ?NoahSlater
Comment: "...definitely a problem if they are not required". Hm. following that analogy it could be applied to these too: "Upstream-Name (optional), Upstream-Maintainer (optional), Upstream-Source (optional)". Optional fields are optional by defnition as per RFC2822 that has been set as an example. Complexity where? All fields here are easy to type, understandable at a glance, machine parseable and extendable. "not useful within the context"? The solution offered, single URL in Upstream-Source to a web page or tarball does not cover the case where there is no tarball or web page to point to (e.g. snapshots packaged directly from VCS repositories; when there is no upstream any more, but someone has put the code to VCS etc.). If the copyright file is required to point to the location of the sources, there should be no assumptions of the form of the URI. A good design allows current and future use. Field definitions that are non-extendable -- that work in some cases, but not others, that lack expressive power for details -- need an improvement. This is comparable to information that is attached to the licences after "BY" keyword (see below EBNF). To give an example: at the time of writing 2008-11-08, your Debian package [http://packages.debian.org/changelogs/pool/main/p/planet-venus/planet-venus_0~bzr95-2/planet-venus.copyright planet-venus] is packaged directly from bzr with version number 0~bzr95-2. The planet-venus package's debian/copyright did not point to a upstream tarball, but to VCS browser via field Upstream-Source. The URL mentined cannot be used alone to retrieve the sources in any machine automated way. It needs further specifications. By using fields Upstream-Vcs-Browser: http://www.intertwingly.net/code/venus/ and Upstream-Vcs-URI: type=bzr; uri=http://intertwingly.net/code/venus, these would have siganalled enough information for a tool to select between show or proper retrieval method for the sources. -- JariAalto
disclaimer for non-free and contrib packages (see [http://www.debian.org/doc/debian-policy/ch-docs.html#s-copyrightfile Policy 12.5])
- only once
- optional
Suggested name: Disclaimer
- Suggested format: free content explaining that the contrib or non-free package is not part of the Debian GNU/Linux distribution and briefly explain why.
Comment: The Developers Reference Section [http://www.debian.org/doc/developers-reference/best-pkging-practices.html#bpp-origtargz 6.7.8.2] requires that debian/copyright contain additional information about repackaged .orig.tar.gz files. This might, for example, be a statement to the effect that all precompiled java class files were removed (this is pretty common!). All the Original-Source-* fields have now been removed though. So where does this fit into these machine-readable copyright files? -- StuartPrescott
Examples:
Format-Specification: http://wiki.debian.org/Proposals/CopyrightFormat?action=recall&rev=196 Upstream-Name: SOFTware Upstream-Maintainer: John Doe <john.doe@example.com> Upstream-Source: http://www.example.com/software/project Upstream-Vcs-URI: type=git; uri=http://git.example.com/project.git Format-Specification: http://wiki.debian.org/Proposals/CopyrightFormat?action=recall&rev=196 Upstream-Name: xyz Upstream-Maintainer: Jane Smith <jane.smith@example.com> Upstream-Vcs-Browser: http://www.example.com/gitwww Upstream-Vcs-URI: type=git; uri=http://git.example.com/xyz.git
Files Section (Repeatable)
The declaration of copyright and license for files is done in one or more stanzas, formatted as RFC2822-style fields. Each stanza is separated from others by a blank line.
- List of files sharing copyright holders and licensing terms
Repeatable field, like the Package field in debian/control
Suggested name: Files
- Suggested format: list of files or patterns, see "File patterns" below
Good place to add mandatory hints on original package authors: use Files: debian/*
Copyright holders for the files listed in the previous Files field
- Repeatable field, one field per copyright statement
Suggested name: Copyright
- Suggested format: a single valid copyright statement
Example Format: Copyright 2008, Holder <email@example.org>
- A valid copyright statement should include the word "Copyright", the copyright year(s) and the copyright holders name. Optional contact information or URI is recommended.
History of package maintainers can optionally be reflected as multiple copyright fields for Files: debian/*
Licensing terms for the files listed in the previous Files field
Suggested name: License
- Suggested format:
First line: See Keywords section.
- Remaining lines: One of the following:
- If a single license keyword is used you must either:
- Copy the full text of the license; or
Leave this section empty if a standalone License section exists (see the Standalone License Section section) that matches the license keyword.
If multiple licenses apply you must not copy any text into the remaining lines. A standalone License section must be used for each license keyword mentioned.
- If a single license keyword is used you must either:
Example:
Files: * Copyright: Copyright 2008, John Doe <john.doe@example.com> Copyright: Copyright 2007, Jane Smith <jane.smith@example.com> License: PSF-2 [LICENSE TEXT]
Discussion
Comment: There initial proposal suggested Copyright: <"Copyright" | "©" | "Copr."> .... The Header field itself expresses this, so a simple format suffices. Note that mere "(C)" is legally null.
Comment: I really wish you would request comments before making changes like this, I am getting rather frustrated. The Copyright field should include whatever copyright statement the original copyright holders have chosen, they certainly do not need to provide a personal name, let alone in the rather ridged and internationally insensitive "<Firstname> <Lastname>" format. The value of this field is to be reproduced verbatim so it needs to include the word "Copyright" or similar, even if this looks to be duplication to you.
Comment: In [http://www.gnu.org/licenses/gpl-howto.html FSF's instructions] abbreviations are discouraged practise. To my knowledge there are no internationally valid abbreviations for word Copyright. Please cite a source that clarify legal status of "Copr." etc. abbreviations. --JariAalto
Comment: Copyright is different from license information. License needs to be "reproduced verbatim", but no such thing applies for Copyright. The Copyright information in source files can vary in format and layout from file to file even in same sources. The names of the authors can be spelled differently in the sources (email addresses, nick names, VCS names, login names). An example: there are files whose Copyright information field lists login names -- say "tobran" (made up here) -- whose real name can be found after looking at other files or Googling from project site, mailing lists. It's perfectly valid to record person's real name to the debian/copyright file's field Copyright instead of the login name -- which cannot serve for claims of Copyright. People own Copyrights, login names and nick names don't. The years mentioned also differ from file to file: editing spans over period of time. There is no need to record files individually and their Copyright years individually. Therefore it makes no sense to require any "verbatim" factor. The important things is that debian/copyright records -- or collects -- the information in coherent manner to a single place for Debian maintainers, other readers (the file is shown at packages.debian.org) and ftp-admins to examine. The "verbatim" information can be found from the sources itself. As far as I can tell, there is no legal requirement to preserve the Copyright text "verbatim"; this can be seen in practise, because that hasn't been required to date in any of the existing debian/copyright files. A citation is needed from a legal source for alledged verbatim requirement. -- JariAalto
Resolved issues
Comment: There is currently a [http://lists.debian.org/debian-devel-announce/2006/03/msg00023.html rather definite request (requirement?)] that copyright files include not just an assertion that the code is under a particular licence but to include the text that states that this is the case (e.g. the standard GPL statement used by many "This program is free software ... You should have received ..."). It appears that this is inconsistent with the the currently proposed format. Are the ftp-masters happy with the format as described here and what amounts to the removal of this statement from the copyright file? Or is it envisaged that the text after the "License:" include this statement? If this is the case, then some clarification is required and the examples should do this. -- StuartPrescott
Comment: That email is over two years old now. I can only assume, given the large number of people and packages using this copyright format as it stands, that the FTP Masters are happy with it to date. -- ?NoahSlater
Standalone License Section
Where a set of files are dual (tri, etc) licensed you must use a single line License field and use standlone License fields to expand the license keywords.
Files: src/js/editline/* Copyright: Copyright 1993, John Doe Copyright: Copyright 1993, Joe Average License: MPL-1.1 | GPL-2 | LGPL-2.1 License: MPL-1.1 [LICENSE TEXT] License: GPL-2 [LICENSE TEXT] License: LGPL-2.1 [LICENSE TEXT]
Where multiple sets of files use the same license you can avoid repetition by using a single line License field and use a separate standalone License field to expand the license keyword.
Files: src/js/editline/* Copyright: Copyright 1993, John Doe Copyright: Copyright 1993, Joe Average License: MPL-1.1 Files: src/js/fdlibm/* Copyright: Copyright 1993, J-Random Comporation License: MPL-1.1 License: MPL-1.1 [LICENSE TEXT]
License Aliases
If a common type of license is a combination of multiple licenses (like the perl license), an alias can be made, so that it can be clear that it's the particular combination of licenses and not just any combination.
Files: * License: Perl License-Alias: Perl Licenses: GPL-2+ | Artistic-2.0 License: GPL-2+ [LICENSE TEXT] License: Artistic-2.0 [LICENSE TEXT]
Discussion
Comment: This seems overly complex to me, lets keep it simple. Using "GPL-2+ | Artistic-2.0" for Perl licensing is fine and should cause no confusion. Adding an extra field to save 13 characters on the odd occasion that the Perl style licensing is chosen is a bad decision. -- ?NoahSlater
Comment: It doesn't have anything to do with saving characters, but to declare, that it's infact, per example, the "perl" license. This could be extended to inlcude the exact copyright note that the upstream had placed while delegating it to two licenses.
License-Alias: Perl Licenses: GPL-2+ | Artistic-2.0 It is free software; you can redistribute it and/or modify it under the terms of either: . a) the GNU General Public License as published by the Free Software Foundation; either version 1, or (at your option) any later version, or . b) the "Artistic License".
I did though of other names, "Multi-License", "Merged-License" etc... but though "License-Alias" was sufficient. -- ?AzaToth
Comment: I disagree, this isn't the "Perl license" - it's the combination of licenses that Perl uses. Just like the Mozilla tri-licensed MPL|GPL|LGPL stuff, I don't see any reason why this needs to be given a distinct name. -- ?NoahSlater
Fields Detail
Files
Format
The value of the Files field should be a list of comma-separated values:
Files: foo.c, bar.*, baz.[ch]
File names containing spaces or commas should be put within double quotes. The backslash character is an escaping character, be it inside or outside double quotes:
Files: "Program Files/*", manual\[english\].txt
Syntax
Patterns are the ones recognised by the find utility's -name and -wholename flags. They behave as if find had been called in the following way from the top source directory:
find . -wholename "$PATTERN"
This will match all Makefile.am files in the tree and all Python scripts:
Files: */Makefile.am, *.py
But this will only match the top-level Makefile.am:
Files: ./Makefile.am
Special rule: if a pattern $PATTERN does not match any file in the source, it is implicitly considered to be expanded to */$PATTERN. This is to avoid insane verbosity when referring to a unique file buried deep in the tree.
Match Order
It is quite common for a work to have files with copyright held by different parties and received under different licenses. To allow this, multiple stanzas are allowed with different Files declarations.
However it makes for easier reading if the copyright file lists the "main" license first: the one matching the "top level" of the work, with others listed as exceptions. To allow this, the following precedence rule applies for matching files: If multiple Files declarations match the same file, then only the last match counts.
As a result, it is recommended for clarity that the stanzas appear in order from most general (e.g. Files: *) first, through to most specific. In the following example, the file getopt.c matches both Files: * and Files: getopt.*; only the last match counts, so the file getopt.c has the license declaration License: BSD-C2.
{{{contain arbitrary values Files: * Copyright: Copyright 2003-2005, John Doe <jdoe@xample.com> License: [the main work's license]
- [LICENSE TEXT]
Files: getopt.* Copyright: Copyright 2000, The Corporation Foundation, Inc. License: BSD-C2
- [LICENSE TEXT]
Files: debian/* Copyright: Copyright [years], [the debian package copyright holder] License: [the debian package license]
- [LICENSE TEXT]
}}}
It is very common for the Debian packaging work to have a different copyright holder and/or license from the upstream work. In these cases, it is important that the debian/* pattern is placed after any other conflicting patterns.
License
Keywords
The "License" field, to be machine-parseable, should not contain arbitrary values. There needs to be a list of accepted keywords which have a very specific, unambiguous meaning. The convention for license abbreviations keyword is XYZ for a license with only one version, and XYZ-n for *sequential version n* of the XYZ license, where XYZ-1 is chronologically earlier than XYZ-2. The syntax of the License keyword can be defined using similar to [http://en.wikipedia.org/wiki/Extended_Backus–Naur_form EBNF grammar]:
License ::= <keyword>[<version>]["BY" {N* <clarification>}] <keyword> ::= <License keyword FSF> | <License keyword BSD> | <License keyword well known> | "other" <License keyword FSF> ::= GPL | LGPL | AGPL | GFDL <License keyword BSD> ::= ... to be decided, see table below for proposals. <License keyword well known> ::= ... to be decided, see table below for proposals. <version> ::= <License version>["+"] <License version> ::= <Numeric version> | <Other version> <Numeric version> ::= [0-9.]+ <Other version> ::= <Vendor's version> | <publication date> <Vendor's version> ::= string (* anything: "A", "B", "public" *) <publication date> ::= YYYYMMDD <clarification> ::= <clarification GPL> <clarification GFDL> | <clarification CC> | <clarification other> <clarification GPL> ::= "-CC" (* With Creative Commons' metadata and Commons Deed added to GPL *) "-OpenSSL" (* With openSSL exception added *) <clarification GFDL> ::= "-NIV" (* With no invariant sections *) "-CC" (* With Creative Commons' metadata and Commons Deed added to GFDL *) <clarification CC> ::= (* Creative Commons License variants *) "-NC" (* No Commerercial *) "-ND" (* No Derivative Works *) "-SA" (* Share Alike *) <clarification other> ::= Other list of license specific keywords that clarify optional parts included or exluded.
Examples of the above ENBF:
keyword |
meaning |
Apache-1.0 |
Apache license, version 1.0 only |
CC-3.0-BY-SA-ND-NC |
Creative Commons Attribution License 3.0; with Share Alike, No Derivs, No commercial |
GPL-2+ |
GNU General Public License, version 2 or (at your option) any later version |
GPL-3 |
GNU General Public License, version 3 only |
LGPL-2 |
GNU Lesser General Public License, version 2 only |
LGPL-2+-BY-OpenSSL |
GNU Lesser Public License, version 2 or (at your option) any later and with OpenSSL linkage exception added |
GFDL-1.2 |
GNU Free Documentation License, version 1.2 only |
GFDL-1.2+-BY-NIV |
GNU Free Documentation License, version 1.2 or (at your option) any later, with no invariant sections |
PSF-2 |
Python Software License, version 2 only |
other |
Any other custom license. Text must be copied verbatim. |
Notes:
The extra plus-sign in <License version>["+"] means that the License contains clauses similar to GPL's "version N or (at your option) any later version".
The GPL/LGPL BY-OpenSSL clarification gives permission in cases where GPL code is linked with OpenSSL library. For more information, see page [http://www.gnome.org/~markmc/openssl-and-the-gpl.html "The OpenSSL License and The GPL"] by Mark ?McLoughlin and message [http://lists.debian.org/debian-legal/2004/05/msg00595.html "middleman software license conflicts with OpenSSL"] by Mark ?McLoughlin in Debian legal mainling list. The License needs to include following or close to similar excerpt to qualify for BY-OpenSSL which allows linking with OpenSSL:
In addition, as a special exception, the copyright holders give permission to link the code of portions of this program with the OpenSSL library under certain conditions as described in each individual source file, and distribute linked combinations including the two. You must obey the GNU General Public License in all respects for all of the code used other than OpenSSL. If you modify file(s) with this exception, you may extend this exception to your version of the file(s), but you are not obligated to do so. If you do not wish to do so, delete this exception statement from your version. If you delete this exception statement from all source files in the program, then also delete it here.
Discussion
Comment: I am opposed this suggestion. I think introducing an EBNF grammar for the license keywords is totally unnecessary. This seems like complexity for complexities sake. No one is suggesting that this field contain arbitrary values, the original proposal was that the values would be explicitly enumerated. There is no technical reason why this is not sufficient for our needs. In the edit to add this proposal you have completely obliterated the existing list of keywords that were already within use in Debian. -- ?NoahSlater
Comment: On the contrary. There is no need to list every possible case starting from GPL-1, GPL-1+, GPL-2, GPL-2+. LGPL-1...LGPL-2.1, LGPL-2.1+ etc. The grammar can sufficiently express how the license keywords are to be constructed. If the examples above are not sufficient, they can be extended as needed. The EBNF does not imply arbitrary values. What needs to be decided is the exact keywords that express the license names for slots "to be decided". -- JariAalto
Comment: You say "there is no need", but this doesn't really make sense. We have two suggestions so far, enumerating every single licence using, essentially opaque, identifiers, or using your solution and making the identifiers themselves meaningful, constructing a grammar to do so. Two solutions, and varying positive and negative points. I prefer the first, and original, solution of explicitly enumerating each license, I find the alternate to be overcomplex, and buys us little for the additional complexity it adds for both implementors of this standard and those trying to produce conforming documents.
Comment: The grammar can "enumerate" licenses because that's what EBNF does. This does does not exclude in any way that there could be a "whole list" if someone wished to write one with every possible combination. The idea is that when more and more licenses are added, there is a need to define syntax how the license keywords are to be constructed. -- JariAalto
Comment: We had already written a whole list that you have subsequently removed. This list was already in use by a large number of Debian packages. Could you please restore it? -- ?NoahSlater
Comment: That list is expressed in the grammar; the examples section demonstrates how to apply it. No license keywords should have been lost. For example you can read: GPL-1, GPL-1+, GPL-2, GPL-2+, GPL-3, GPL-3+ etc. which are all covered. -- JariAalto
Comment: Another common GPL exception is [http://www.gnu.org/licenses/gpl-faq.html#FontException Font Exception]. Should it be added to the above EBNF? -- TheppitakKaroonboonyanan
<License keyword BSD>
Proposal 1:
Comment: These licenses have no recognised version number. Perhaps the abbreviation should deliberately show the number of clauses in a way that doesn't indicate a version number: BSD-C<N>. But that still sounds vaguely like a version number; it still suggests sequence in the different versions that doesn't actually match the true chronology; it also fails to indicate *which* clauses are included. —BenFinney 2008-10-16:
BSD-C2 |
Two-clause BSD license |
BSD-C3 |
Three-clause BSD license, with no-endorsement clause, as seen in /usr/share/common-licenses/BSD?BR |
BSD-C4 |
Four-clause BSD license, with no-endorsement clause and advertising clause; GPL-incompatible (need exact text) |
Proposal 2:
In the absence of a clear succession of differently-numbered consecutive versions of a license text, my proposal is: we could come up with abbreviations similar to those used for indicating active clauses in the Creative Commons licenses. This way, no false impression of chronological sequence is implied, and the abbreviation provides a mnemonic for what the terms of the license actually are, not just the number of clauses in them. -- BenFinney 2008-10-15
BSD-BY-LC |
forms requiring only the inclusion of copyright notice and condition |
BSD-BY-LC-NE |
plus “no endorsement without permission” |
BSD-BY-LC-NE-AD |
plus “advertising required” |
Proposal 3:
While mnemonics like LC,NE,AD may sound practical (where does LC abbreviation come from?), in BSD case I find them lacking wider recognition. The BSD licenses have been examined by FSF, so perhaps we could use names and definitions used at [http://www.gnu.org/philosophy/license-list.html#FreeBSD License list] —JariAalto:
keyword |
GPL compatible |
meaning |
[http://www.gnu.org/philosophy/license-list.html#OriginalBSD BSD] |
No |
Also known as the “[http://www.opensource.org/licenses/bsd-license.php Original 4-clause BSD license]”. Contains the “obnoxious BSD advertising clause”. |
[http://www.gnu.org/philosophy/license-list.html#ModifiedBSD ModifiedBSD] |
Yes |
This is the original BSD license, modified by removal of the advertising clause. It is a simple, permissive non-copyleft free software license. |
[http://www.gnu.org/philosophy/license-list.html#FreeBSD FreeBSD] |
Yes |
Also known as the “[http://www.freebsd.org/copyright/freebsd-license.html 2-clause BSD license]”. Original BSD license with the advertising clause and another clause removed. Simple, permissive non-copyleft free software license. |
[http://www.gnu.org/philosophy/license-list.html#ISC OpenBSD] |
Yes |
Also known by name "[http://www.opensource.org/licenses/isc-license.txt ISC License]". This license does have an unfortunate wording choice. |
[http://www.gnu.org/philosophy/license-list.html#clearbsd ClearBSD] |
Yes |
Based on the modified BSD license, and adds a term expressly stating it does not grant you any patent licenses. |
<License keyword well known>
The basic of these keywords should be wide recognition. Something based on lists like [http://www.opensource.org/licenses/alphabetical Open Source Initiative: Licenses by Name], [http://en.wikipedia.org/wiki/List_of_FSF_approved_software_licences Wikipedia: List of FSF approved software licences], [http://www.gnu.org/philosophy/license-list.html FSF: Various Licenses and Comments about Them] and [http://www.opensource.org/licenses/alphabetical Open Source Initiative: Licenses by Name]
Comment: What do you mean by "wide recognition", I hardly see how this has any relevance. If we need to use a license, we need to use a license. -- ?NoahSlater
Comment: E.g. using label Apache to refer to particular type of license is widely recognized (cf. FSF, Open Source initiative). This list should include only licenses that are in wide use and refrain from listing all possible past and present licenses. The lesser known licenses can use category "other" with text copied verbatim. The list below is not yet complete. -- JariAalto
TODO: This list needs better scrutiny:
Comment: What do you mean better scrutiny? For what? -- ?NoahSlater
Comment: This list does not yet represent a solid list of known licenses. That should probably be determined by a wider audience (poll in Debian Legal, Devel?). Good candicates are those listed in FSF or OSS pages mentioned above. -- JariAalto
keyword |
meaning |
Apache |
Apache license |
Artistic |
The Perl Artistic license. Always spcify the version number, either [http://www.perlfoundation.org/artistic_license_1_0 1.0] or [http://www.perlfoundation.org/artistic_license_2_0 2.0] |
CC |
[http://creativecommons.org/license Creative Commons Attribution License] |
IBMCPL |
IBM Common Public License |
CeCILL |
CEA-CNRS-INRIA-Logiciel Libre. See http://www.cecill.info/licences |
Eiffel |
Eiffel Forum License |
Erlang |
Erlang Public License |
Expat |
The terms of the Expat license, http://www.jclark.com/xml/copying.txt ?BR This license is what many people mean by "the MIT license", but that term is too ambiguous as there is more than one "MIT license" in the wild |
ISC |
The Internet Software Consortium's “ISC license”. See http://opensource.org/licenses/isc-license.txt |
LatexPPL |
The LaTeX Project Public License. See e.g. http://www.latex-project.org/lppl/lppl-1-3a.txt; GPL-incompatible?BRNote that works under any version of the License often have additional restrictions attached; check carefully. |
MPL |
Mozilla Public License. See e.g. http://www.mozilla.org/MPL/MPL-1.1.html |
PSF |
Python License |
PHP |
PHP License |
W3C-Software |
The W3C Software License. See http://www.w3.org/Consortium/Legal/2002/copyright-software-20021231 |
ZLIB |
The zlib/libpng license. See http://www.opensource.org/licenses/zlib-license.php |
Zope |
Zope Public License. |
Things than need to be clarified
If author did not specify version and version cannot be decoded from the context or other files, how the license should be expressed?
MIT |
Several variants of the MIT license exist: (1) the standard version with three paragraphs (blanket permission, keep this notice, NO WARRANTY), (2) a version with a no-endorsement clause, and (3) other versions with slight wording differences. Text needs to be copied verbatim |
PD |
[http://creativecommons.org/licenses/publicdomain/ In the public domain]. Not applicable everywhere in the world. Text needs to be copied verbatim |
Comment: Can we have other-non-free and other-gpl[2|3]-[in]compatible? The latter would help to automate a GPL-compatibility check. The former could help with mixed free/non-free source packages -- the whole package would of course go in non-free, but it could be worth noting for a later effort to separate out the non-free parts. -- ?AdamPowell
Syntax
License names are case-insensitive.
The value of the field should follow the syntax of debian/control's Depends field. The pipe character "|" is used for code that can be used under the terms of either licenses. The comma "," is used for code that must be used under the terms of both licenses (for rare cases where a single file contains code under both licenses).
For instance, this is a simple, "GPL version 2 or later" field:
License: GPL-2+
This is a dual-licensed GPL/Artistic work such as Perl:
License: GPL-2+ | Artistic-2.0
This is for a file that has both GPL and classic BSD code in it:
License: GPL-2+, BSD-C2
And this is for a file that has Perl code and classic BSD code in it:
License: GPL-2+ | Artistic-2.0, BSD-C3
A GPL-2+ work with the OpenSSL exception is in effect a dual-licensed work that can be redistributed either under the GPL-2+, or under the GPL-2+ with the OpenSSL exception. It is thus expressed as GPL-2+-BY-OpenSSL:
License: GPL-2+-BY-OpenSSL In addition, as a special exception, the author of this program gives permission to link the code of its release with the OpenSSL project's "OpenSSL" library (or with modified versions of it that use the same license as the "OpenSSL" library), and distribute the linked executables. You must obey the GNU General Public License in all respects for all of the code used other than "OpenSSL". If you modify this file, you may extend this exception to your version of the file, but you are not obligated to do so. If you do not wish to do so, delete this exception statement from your version."
Discussion
Resolved issues
2008-11-21: The following below has been resolved by the use of EBNF grammar. The GPL/LGPL extensions are defined after "BY" keyword. In this case the -OpenSSL --JariAalto
The syntax of this example is not consistent with the syntax description above. The use of GPL-2+ | other is only permissible in a Files section and not in a stand-alone License section. The inclusion of free-form text in the remaining lines of the Licence field of a Files section is not permitted if more than one license keyword is being used. As currently written in this proposal, this needs to be:
Files: * Copyright: Copyright 1999-2007 John Doe License: GPL-2+ | other License: GPL-2+ On Debian systems, [...] License: other In addition, as a special exception, [...]
Comment: However, this description is still not what the author meant. If the field read License: GPL-2 | Artistic-2.0 then the end-user has the choice of choosing to accept either the GPL-2 or Artistic 2.0 licenses, but this is not the case and this "other" is not really a separate license as it is instead an exception that is added to notice putting the code under the GPL in the first place. There is no choice between "GPL-2+" and "other" as indicated by the use of the | between the license keywords. The problem stems from conflating the statement of what licenses the code is under with the licenses themselves.
Comment: I disagree: It was indeed the intention of the author to license as either "GPL2+" or "GPL2+ adjusted to please OpenSSL", which is 2 separate distinct licensing terms, even if written more compact to not repeat the GPL2+ part. In other words, I believe the above is correct. -- JonasSmedegaard
Comment: You are right in that license can be reduced to "GPL-2" | "GPL-2 with exception" because the exception is separable for derived works. But expressing "other" as being purely the exception is wrong as it is not a self-contained licence. The way the example is written above illustrates this... the field starts with the words "In addition" and the reader is left wondering "in addition to what?". The human-readable aspect of the copyright file is lost if such ambiguities are left in it and the machine-parsable aspect is also lost if some other licenses are really a self-contained license while others are actually an addition to another section elsewhere in the document. However, this may be the way of solving these problems: can it be expressed as GPL-2+ | GPL-2+ with exceptions where the License stanza for "GPL-2+ with exceptions" states that it is "GPL-2 In addition, as a special exception..."? It would presumably be desirable to make "GPL-2+ with exceptions" into something that doesn't contain whitespace to be consistent with the other keywords though -- StuartPrescott
Comment: Does it instead need to be written explicitly like this:
Files: * Copyright: Copyright 1999-2007 John Doe License: GPL-2+ + exception On Debian systems, [...] Exception: In addition, as a special exception, [...]
Comment: Notes:
I considered GPL-2+, exception, but I think it's important to distinguish between the exception being a separate license and it modifying the previously listed licence hence "+" seems better than "," which already has another meaning in this field.
Perhaps resurrecting the License-Terms field would be better than creating an Exception field for this.
- This is related to the earlier question about the removal of the actual statement of how the code was released from the copyright file (e.g. the "This program is free software ... You should have received ..." text). This openssl exception is part of that statement and not part of a license.
-- StuartPrescott
Comment: I do not like the "+" syntax or the "Exception" field. I think this should be indicated with the "other" keyword exlusively, with an explanation. It is misleading to claim this is still GPL. -- ?NoahSlater
Comment: To do so would mean that, as JonasSmedegaard noted, we lose the information that GPL+Exception == GPL | GPL+Exception. So while saying "GPL+Exception" == "other" has merits, it would throw away a lot of information from this copyright format, especially since (as noted above) derived works need not follow the exception. Since GPL+Exception are common enough terms for permitting code to be distributed, it would be worth having a syntax that actually allows us to express these terms both accurately and succinctly. Perhaps writing it explicitly as GPL | GPL+Exception rather than just GPL | other is both accurate and self-contained. -- StuartPrescott
My only issue is with the use of spaces in the license keywords, how about: GPL/extra, GPL2/extra, GPL2+/extra, etc or similar?
Examples
Simple
A possible copyright file for xsol:
Format-Specification: http://wiki.debian.org/Proposals/CopyrightFormat?action=recall&rev=143 Upstream-Name: X Solitaire Upstream-Source: ftp://ftp.example.com/pub/games Files: * Copyright: Copyright 1998, John Doe <jdoe@example.com> License: GPL-2+ On Debian systems the full text of the GNU General Public License can be found in the `/usr/share/common-licenses/GPL' file. Files: debian/* Copyright: Copyright 1998, Jane Smith <jsmith@example.net> License: other [LICENSE TEXT]
Complex
A possible copyright file for planet-venus:
Format-Specification: http://wiki.debian.org/Proposals/CopyrightFormat?action=recall&rev=178 Upstream-Name: Planet Venus Upstream-Maintainer: John Doe <jdoe@example.com> Upstream-Source: http://www.example.com/code/venus Files: * Copyright: Copyright 2008, John Doe <jdoe@example.com> Copyright: Copyright 2007, Jane Smith <jsmith@example.org> Copyright: Copyright 2007, Joe Average <joe@example.org> Copyright: Copyright 2007, J. Random User <jr@users.example.com> License: PSF-2 [LICENSE TEXT] Files: debian/* Copyright: Copyright 2008, Dan Developer <dan@debian.example.com> License: GAP Copying and distribution of this package, with or without modification, are permitted in any medium without royalty provided the copyright notice and this notice are preserved. Files: debian/patches/theme-diveintomark.patch Copyright: Copyright 2008, Joe Hacker <hack@example.org> License: GPL-2+ [LICENSE TEXT] Files: planet/vendor/compat_logging/* Copyright: Copyright 2002, Mark Smith <msmith@example.org> License: MIT [LICENSE TEXT] Files: planet/vendor/httplib2/* Copyright: Copyright 2006, John Brown <brown@example.org> License: other Unspecified MIT style license. Files: planet/vendor/feedparser.py Copyright: Copyright 2007, Mike Smith <mike@example.org> License: PSF-2 [LICENSE TEXT] Files: planet/vendor/htmltmpl.py Copyright: Copyright 2004, Thomas Brown <coder@example.org> License: GPL-2+ On Debian systems the full text of the GNU General Public License can be found in the `/usr/share/common-licenses/GPL' file.
Questions
Question: I am not quite sure this is the right place... Some licenses must be presented to the *user* that install the system (like sun-java, AFAIK), and forbid preseeding. Some other licenses simply must be presented(like [http://intellinuxwireless.org/?n=faq&s=license#license_1 intel ipw2200] license). Some other licenses should be presented and allow special case for automated installation. A filed like LicenseAcceptation could then be used in preinst/postinst script. FranklinPiat
Recent Changes
2008-11-21: Add to EBNF new topic Notes:. Add OpenSSL exception excerpt and cite Mark ?McLoughlin.
2008-11-20: Add to EBNF new "-OpenSSL" clarification for GPL, LGPL. Add version number to Artistic licenses; also add URLs to point to 1.0 and 2.0. Change obsolete GPL-1+ examples to more in use GPL-2+ examples. Change other-BSD keywords to proposed BSD-C2 keywords in examples. Add new new section Resolved issues
2008-11-08: Simplify Upstream-Vcs-URI. Use example.{org,com,net} in all URLs. —JariAalto
2008-11-07: License Keywords: Write generic EBNF. Divide BSD cases to proposals 1,2,3. Make generic keyword tables. Add BSD table for proposal 2. Add Comment: indicators. Add Discussion sections. Properly indent all comments. (MAJOR CHANGES). —JariAalto
2008-11-02: License Keywords: Add table about FSF recognized names for use with BSD licenses. —JariAalto
2008-10-21: Header Section: Add Upstream-Vcs-Browser and Upstream-Vcs-URI. —JariAalto
2008-10-15: Proposal for abbreviations that indicate *which* BSD-style license is referenced. —BenFinney
2008-09-27: Re-work some of the provisional language, as this is shaping up more as a specification to be recommended (not there yet, but the document wording should get ready for this). —BenFinney
2008-09-27: Clarify the "match order" explanation and example, using positive rather than negative examples; negative examples proved confusing to multiple readers. —BenFinney
2008-07-25: Show that "©" is valid for copyright statement but use "Copyright" in most examples, as the latter is easier for most people to type. -- ?NoahSlater, BenFinney
2008-07-25: Incorporate suggestion of repeatable Copyright field, as no significant drawbacks have been raised. —BenFinney
2008-07-22: Note line-length problems with long Format-Specification URLs, and fix examples to avoid lintian warnings in packages that use the examples. —BenFinney
2008-07-12: I propose a new header field, Disclaimer, for contrib and non-free package to comply with the suggestion introduced by Policy 3.8.0 to state that they are not part of Debian. -- CharlesPlessy
2008-07-11: Improved explanation around the License field. -- ?NoahSlater
2008-07-11: Removed the mention of License-Terms field following no reply to criticisms. -- ?NoahSlater
2008-06-14: Moved discussion of broader change to date+time format to ["Proposals/DatetimeFormat"] -- BenFinney
2008-05-28: Changed Packaged-Date time format to ISO 8601. Corrected examples as well. -- ?TeemuIkonen
2008-05-26: Added Upstream-Maintainer field, which should point to the current upstream contact (some packages have retired authors which should be named in the Upstream-Author field, but are not active anymore) -- ?TeemuIkonen
2008-05-19: Implement ?NoahSlater's suggested change, "Can we take this oportunity to use the word "packaged" instead of "debianized"?
2008-05-02: Removed the section on Notice because I don't see a use case for this. Any licenses in /usr/share/common-licenses should be expressed using a regular License field with the standard Debian notice text being used as the long description. Unknown or uncommon licenses can be expressed using the "other" license type with the long description serving as normal.
2008-05-01: Once more: reorganisation (by sections, by fields) -- ?MathieuParent
2008-04-29: Added TOC, recent changes at bottom, separating sections with empty lines is mandatory to parse "Multiple license fields" -- ?MathieuParent
2008-04-28: Added clarified multiple License field behaviour. -- ?NoahSlater
2008-04-28: Added back Original-Source-* style header fields as these are required for packages where the upstream source is taken directly from a repository or an unversioned tarball. -- ?NoahSlater
2008-04-28: Simplify, make it rfc2822 compatible, disambiguation (I've changed/removed stuff from other people, please don't be upset !, old is in ["Proposals/CopyrightFormat/Archive"] ) -- ?MathieuParent
- 2008-04-01: Added proposal to use License independently to avoid repetition.
2008-03-30: Added GFDL 1.1 and 1.1+ -- ?GustavoMontesino
2008-03-26: Added CC-BY-SA-3 license. -- ?NoritadaKobayashi
2008-03-19: Proposed X-Non-Free-Autobuild field for non-free packages -- CharlesPlessy
Comment: Today's format is XS-Autobuild: yes as announced [http://lists.debian.org/debian-devel-announce/2006/11/msg00012.html here] -- SteffenMoeller
2008-02-29: Added EFL-2 license. -- ?NoahSlater
- 2008-02-14: Added ZLIB license (the zlib/libpng one). -- Kaeso
2008-01-12: Added section describing proposal for full RFC822 specfication. -- ?MortenKjeldgaard, Discussion on -devel http://lists.debian.org/debian-devel/2007/08/msg00242.html
2007-10-20: Added CC-BY-3 license. -- ?NoahSlater
2007-10-09: Added proposal for all-permissive licenses. -- ?NoahSlater
2007-10-09: Added the MPL-1.1 and EPL-1.1 licenses. -- ?NoahSlater
2007-08-09: replaced "GPLvX" with "GPL-X" to match /usr/share/common-licenses filenames. Replaced "BSD" with "BSD-3". --Sam
2007-08-09: reverted the parentheses proposal for /usr/share/common-licenses, it made lines too long. --Sam
- 2007-08-06: add Artistic-2.0. -- Josh Triplett
- 2007-08-06: added MIT licenses to the "Stuff that we might want" section; needs disambiguation. -- Josh Triplett
- 2007-08-06: clarified "Artistic" with a specific reference to the original license as found in common-licenses. -- Josh Triplett
- 2007-08-06: added several more licenses; suggested changing BSD to BSD3; refer to the Library or Lesser GPL as appropriate to the version. -- Josh Triplett
- 2007-08-05: suggested a way to get rid of the "On Debian GNU/Linux systems, the text of the GPL blahblah" text.
2007-08-05: changed the Files behaviour as suggested by several people on debian-devel.
Aug 2007-08-05: removed Packaging-Copyright and Packaging-License in favour of the Files: debian/* technique. --Sam
2007-08-05: dropped the Source field. It was not really needed for any machine-interpretable work, there's the watch file for that. --Sam