License of Unicode tables (was: Re: Feature request for `describe-char-after')

[Simon Josefsson]

Stefan Monnier <address@hidden> writes:

>> Could someone please explain exactly what it is in the license that
>> make it non-free?
>
> It does not permit modification and distribution of modified versions.
> I.e. it does not allow bug-fixing.

That wasn't clear to me from the license I quoted -- I'll ask if this
really is the intention with the license.  To me, the license said I
could extract information from the files and use in my application.  I
couldn't follow how that prevented me from only incorporating part of
the file, and using something else instead of that part, which
essentially is what a bug fix is.  Has any lawyer given this advice?

But in any case -- is this a realistic problem?  Do we know of any
bugs in the file?  I'd say that changing anything in the file is not
the proper solution, it would only confuse our users since we wouldn't
follow the standard in that case.  The right thing would be to report
this to Unicode Inc. and when they publish a new version of the
tables, we would incorporate it.  If we fork the standard, I think we
get an obligation to our users to document things fully, and also to
stop claiming we follow the "real" Unicode standard; I doubt we want
to go there.

Thanks.

(I have a non-Emacs interest in this question, since GNU Libidn uses
several Unicode Inc. tables.  So even if this is moving off-topic, I'd
still like to clear up the matter.  I suspect Debian would be
interested as well, as they ship several Unicode tables, and IIRC they
have some policies that say what they ship should be free, except for
the explicitly non-free packages.)