Subject: Re: Request for approval: EUPL (European Union Public Licence)
From: "Ben Tilly" <btilly@gmail.com>
Date: Wed, 26 Mar 2008 21:06:16 -0700

On Wed, Mar 26, 2008 at 8:20 PM, Matthew Flaschen
<matthew.flaschen@gatech.edu> wrote:
> Russ Nelson wrote:
>  > Matthew Flaschen writes:
>  >  > Not really, because once an error is found it can be used repeatedly by
>  >  > non-speakers.
>  >
>  > I don't understand your concern here.  If a license in one language
>  > complies with the OSD, AND if the license says "distributors may
>  > choose any translation" THEN people may always distribute under the
>  > license that complies with the OSD.  Thus, the license is
>  > approvable (not that the EUPL says that, but I propose that it
>  > should, to eliminate the translation problem.)
>  >
>  > What are you worried about?  That somebody might have MORE freedom
>  > than the licensor intends?  How is that your concern?
>
>  My concern is the license is stated to be a copyleft license, but an
>  error in a translation could allow circumventing the copyleft.

Copyleft or not has no relevance on the freedoms that are part of the
OSI free software definition.  I grant that it is a valid concern for
software authors considering using the license.  But the OSI is not in
the business of telling people which open source license does what
they want.

>  There are other possibilities too.  I just don't think OSI has the
>  resources to review non-English licenses at this time.

We've not been asked to review non-English licenses.  As far as I'm
concerned, if the English version meets the OSI guidelines, the
license as a whole does to my satisfaction and it should be approved.

The possible question marks are OSD #5, #6 and #10.  In the first two
cases it is possible that a language specific version discriminates
*for* some use or group so strongly that it could be interpreted as
descrimination *against* everyone else.  This does not concern me
because the "discrimination against" can't be significant if the
English version is free enough to be called open source.  #10 is more
tricky, it is possible that there is an option available in another
language that depends upon a specific technology.  However I think
this unlikely given competent translators.  But more critically, the
point of that clause is to forestall a license trap.  If the English
version grants enough freedom that we see no license traps of concern,
then the existence of some technology dependence in some language
version does not constitute a technology trap which is significant.

This all, of course, depends on the *recipient* being able to choose
the license language.  If it is the *distributer* that makes the
choice, then all bets are off and I'd be against accepting any of
these licenses on principle.

Ben