[Bug localedata/11213] localedata licencing issues

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

[Bug localedata/11213] localedata licencing issues

tromey at sourceware dot org
http://sourceware.org/bugzilla/show_bug.cgi?id=11213

--- Comment #13 from Jonathan Nieder <jrnieder at gmail dot com> 2012-07-15 19:05:10 UTC ---
Hi Eben et al,

Petr Baudis wrote:

> I'm sorry, I didn't realize the presence of this bug earlier. We have made some
> research regarding this on the mailing list some time ago and this is currently
> the opinion of FSF on the matter:
>
>   http://sourceware.org/ml/libc-locales/2012-q2/msg00136.html
>
> I.e., locale data is not copyrightable, therefore it cannot be covered by any
> licence and that is also the reason we do not require copyright assignment
> paperwork from the locale authors.
>
> If you (e.g. Debian project) disagree, I encourage you to contact
> [hidden email] or SFLC for further discussions. For us developers,
> this situation really is the best outcome since no paperwork is needed, I'd
> say. My opinion is that the outcome of this bug should be removal of licence
> notices (and "copyright" notices) from all localedata files to clear up any
> possible confusion - comments?

Legal question for you.

Along with everything else it contains, the GNU C library provides a
collection of locale data.  Localedata files include basic information
about how programs should interact with users in a particular area:
currency symbols, paper size, which encoding to use for text files,
translations for "yes" and "no", and other details like that.  You can
find the glibc locales in the localedata/locales directory of glibc:

  http://sourceware.org/git/?p=glibc.git;a=tree;f=localedata/locales

For example, the English locale for the United States is available
from the following address.

 
http://sourceware.org/git/?p=glibc.git;a=blob_plain;f=localedata/locales/en_US

Since this is mostly factual information without much creative
content, no one paid much mind to its license.  Many locales permit
use, distribution, and commercial use without permitting modification:

    # Distribution and use is free, also for
    # commercial purposes.

Most notably, the POSIX locale contains that notice.  Some are more
philosophical:

    % Distribution and use is

Josh Triplett (cc-ed) noticed this in 2009 and reported it to the
Debian project[1].

We are concerned that although these files do not contain much
creative content, they do contain some, for example in their comments.
We are generally not lawyers and do not know what does and doesn't
fall under copyright protection in the United States and elsewhere.
It is important for it to be very clear to both us and our users what
their rights are.

Helge (cc-ed) has taken a survey of authors of locales with the notice
that doesn't permit modification to find what license terms they
intend.  Some findings:

 - many locale authors expected that, as part of glibc, these
   would have the same license as glibc (LGPL-2.1+)

 - they did not intend to forbid modification, and the notice was
   propagated from the POSIX locale by copy and paste.

 - when asked what terms they would like going forward for their work,
   most prefer "public domain", followed by LGPL and GPL.

You can find a summary of Helge's efforts at [2], and the actual
emails are at [1].

Unlike most of glibc, the FSF has not historically required copyright
assignment for these files.  Independently of the above investigation,
they were recently asked about and reaffirmed this position[3]:

    Well that was fast. The SFLC said that this type of thing isn't
copyrightable, and that
    paperwork isn't really necessary. So we should be good to go. Thank you so
much for all your
    help.

Questions:

 - Some locales (namely km_KH, lo_LA, th_TH, and uk_UA) contain
   copyright notices.  Are we legally permitted to remove the
   notices or to change them to say "Authors: ..."?

 - Suppose I wanted to add the text "You may freely use, modify,
   distribute, and relicense this file" to each of the locale data
   files, to make it completely clear that users are free to
   incorporate text from them into differently licensed works.  Would
   that be legally permissible?  Would it be accurate?  Is there any
   reason not to do it?

 - When locale authors have stated a preferred license, is there value
   in documenting that, or would it be counterproductive?

 - If the legal heir to some author of many locales wants to be really
   nasty, what is the worst they can do on the basis of this
   contribution?

Thanks in advance for looking this over.

Sincerely,
Jonathan

[1] http://bugs.debian.org/555168
[2] http://www.helgefjell.de/debianitem.php?name=bug555168
[3] http://sourceware.org/ml/libc-locales/2012-q2/msg00136.html

--
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.
Reply | Threaded
Open this post in threaded view
|

Re: [Bug localedata/11213] localedata licencing issues

Keld Simonsen-2
Hmm, I contributed a number of locales. I always assumed that there was a copyright to each of them.
I do hold a masters in legislate law.

In my work, there was a considerable amount of labour. My initial work was about 100 pages for
the POSIX standard 1003.2 back in 1993.  The expression was quite inventive at the time (IMHO)
and it had a number of facilities, such as being able to be run in many character sets,
it was character set independent - which was a novelty then and interesting, because
we has so many charsets then, and not almost just UTF8 as of today. It was 20 years ago!

Also the machinery had an elaborate set of character names that had had quite some design
work for it to be mnemnoninc. The sorting was an innovative way of sorting almost
all of the characters (in use at that time) which had never been done before.
There was quite some work in getting the data for each country and language, and
also to get interested parties to agree on the data.
The system with locales and  charmaps and repertoiremaps itself was also quite inventive.
And just getting it to compile was also some effort.

But then I attached a licence to it that was not compatible with OSI definitions.
I actually think the locales preceded the OSI defs in time. I wrongly thought that I
should be a kind of maste editor of all of the data, this did not work out.

I am happy with what came out of it, and all the activity and involvement from all over
the world on the glibc localedef data. I recently stated that all of the data I have
contributed is released under GPL v2. I hope this solves some problems.

With my law background I think is is not fair to say the there is no
work height in the data. I don't care too much for my own work wrt. the licences,
My aim was that the data and the work should be used, and also to maintain some quality
to the work. The data should not become flawed. The current scheme with glibc is a
good way to ensure that - but I did not envisage that when I wrote the specs
and the license.

best regards
Keld


On Sun, Jul 15, 2012 at 07:05:10PM +0000, jrnieder at gmail dot com wrote:

> http://sourceware.org/bugzilla/show_bug.cgi?id=11213
>
> --- Comment #13 from Jonathan Nieder <jrnieder at gmail dot com> 2012-07-15 19:05:10 UTC ---
> Hi Eben et al,
>
> Petr Baudis wrote:
>
> > I'm sorry, I didn't realize the presence of this bug earlier. We have made some
> > research regarding this on the mailing list some time ago and this is currently
> > the opinion of FSF on the matter:
> >
> >   http://sourceware.org/ml/libc-locales/2012-q2/msg00136.html
> >
> > I.e., locale data is not copyrightable, therefore it cannot be covered by any
> > licence and that is also the reason we do not require copyright assignment
> > paperwork from the locale authors.
> >
> > If you (e.g. Debian project) disagree, I encourage you to contact
> > [hidden email] or SFLC for further discussions. For us developers,
> > this situation really is the best outcome since no paperwork is needed, I'd
> > say. My opinion is that the outcome of this bug should be removal of licence
> > notices (and "copyright" notices) from all localedata files to clear up any
> > possible confusion - comments?
>
> Legal question for you.
>
> Along with everything else it contains, the GNU C library provides a
> collection of locale data.  Localedata files include basic information
> about how programs should interact with users in a particular area:
> currency symbols, paper size, which encoding to use for text files,
> translations for "yes" and "no", and other details like that.  You can
> find the glibc locales in the localedata/locales directory of glibc:
>
>   http://sourceware.org/git/?p=glibc.git;a=tree;f=localedata/locales
>
> For example, the English locale for the United States is available
> from the following address.
>
>  
> http://sourceware.org/git/?p=glibc.git;a=blob_plain;f=localedata/locales/en_US
>
> Since this is mostly factual information without much creative
> content, no one paid much mind to its license.  Many locales permit
> use, distribution, and commercial use without permitting modification:
>
>     # Distribution and use is free, also for
>     # commercial purposes.
>
> Most notably, the POSIX locale contains that notice.  Some are more
> philosophical:
>
>     % Distribution and use is
>
> Josh Triplett (cc-ed) noticed this in 2009 and reported it to the
> Debian project[1].
>
> We are concerned that although these files do not contain much
> creative content, they do contain some, for example in their comments.
> We are generally not lawyers and do not know what does and doesn't
> fall under copyright protection in the United States and elsewhere.
> It is important for it to be very clear to both us and our users what
> their rights are.
>
> Helge (cc-ed) has taken a survey of authors of locales with the notice
> that doesn't permit modification to find what license terms they
> intend.  Some findings:
>
>  - many locale authors expected that, as part of glibc, these
>    would have the same license as glibc (LGPL-2.1+)
>
>  - they did not intend to forbid modification, and the notice was
>    propagated from the POSIX locale by copy and paste.
>
>  - when asked what terms they would like going forward for their work,
>    most prefer "public domain", followed by LGPL and GPL.
>
> You can find a summary of Helge's efforts at [2], and the actual
> emails are at [1].
>
> Unlike most of glibc, the FSF has not historically required copyright
> assignment for these files.  Independently of the above investigation,
> they were recently asked about and reaffirmed this position[3]:
>
>     Well that was fast. The SFLC said that this type of thing isn't
> copyrightable, and that
>     paperwork isn't really necessary. So we should be good to go. Thank you so
> much for all your
>     help.
>
> Questions:
>
>  - Some locales (namely km_KH, lo_LA, th_TH, and uk_UA) contain
>    copyright notices.  Are we legally permitted to remove the
>    notices or to change them to say "Authors: ..."?
>
>  - Suppose I wanted to add the text "You may freely use, modify,
>    distribute, and relicense this file" to each of the locale data
>    files, to make it completely clear that users are free to
>    incorporate text from them into differently licensed works.  Would
>    that be legally permissible?  Would it be accurate?  Is there any
>    reason not to do it?
>
>  - When locale authors have stated a preferred license, is there value
>    in documenting that, or would it be counterproductive?
>
>  - If the legal heir to some author of many locales wants to be really
>    nasty, what is the worst they can do on the basis of this
>    contribution?
>
> Thanks in advance for looking this over.
>
> Sincerely,
> Jonathan
>
> [1] http://bugs.debian.org/555168
> [2] http://www.helgefjell.de/debianitem.php?name=bug555168
> [3] http://sourceware.org/ml/libc-locales/2012-q2/msg00136.html
>
> --
> Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
> ------- You are receiving this mail because: -------
> You are the assignee for the bug.