Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.

[Brian Thomas Sniffen <bts@alum.mit.edu>]

Sven Luther <sven.luther@wanadoo.fr> writes:

> On Wed, Aug 25, 2004 at 04:40:34PM -0400, Brian Thomas Sniffen wrote:
>> Glenn Maynard <glenn@zewt.org> writes:
>> 
>> > On Wed, Aug 25, 2004 at 09:29:43AM -0400, Brian Thomas Sniffen wrote:
>> >> > The FSF could release a GPL version 3 which has completely arbitrary
>> >> > terms.  If control of the FSF had passed to someone unscrupulous, these
>> >> > terms might be proprietary.  [I'm not saying this is a likely scenario,
>> >> > just a possible one -- I hope this hypothesis seems particularly
>> >> > outrageous.]
>> >
>> > Well, in theory not--"such new versions will be similar in spirit to the
>> > present version".  That vague limitation isn't particularly reassuring,
>> > of course.
>> >
>> >> This is where you lose me.  The FSF releases their GPL v3, which is
>> >> suspiciously similar to a Microsoft EULA.  Now what?  The change I
>> >> submitted, which is distributed with GCC, is licensed only under GPL
>> >> v2.
>> >
>> > Earlier, I wrote a reply asking about things like "v2 vs. v2-or-greater
>> > compatibility" and so on; but after thinking about it for a while, and
>> > rereading the GPL, I realized this is a very common mistaken idea: you
>> > *can not* release your work under "GPL v2, not greater".  GPL#9 says
>> > "if you release under v2, upgrades are allowed".  If you want to release
>> > under v2 without allowing upgrades, you'd have to revoke clause 9--which
>> > would be GPL-incompatible, so you can't do that to your gcc contribution.
>> 
>> We're looking at very different versions of GPL 9.  I'm going to go
>> through it a bit at a time:
>> 
>> > Each version is given a distinguishing version number.  
>> 
>> That's just a statement of fact.
>> 
>> > If the Program specifies a version number of this License which
>> > applies to it and "any later version".....
>> 
>> Well, my changes don't do that, so that's OK.
>
> Notice that one could argue that altough your changes my indeed be licenced
> under GPL v2 and only that one, if you combine it with the real thing, and
> make a binary distribution for example, you are then forced to give the same
> permissions as the original work was under, namely GPL v2 or later.

I don't see any text in the GPL to support that.  The copyleft in the
GPLv2 is in 2b, which says "this License" -- a phrase used often in
the document, and clearly referring to GPLv2 and not any other
license.

That's distinct from GPL 6, which extends the original license to the
public, and forbids further restrictions on the rights to "the
Program" -- that is, the original work, not modifications to it
discussed in GPL 2.

So if I make a change to the Program and distribute my modified
program, combined with the real thing as a binary, as you say, then I
must distribute under the terms of "this License" -- GPL v2.  That's
the only license I received it under.  Though the author may have
offered it to me under a variety of licenses, including the BSD or GPL
v3 licenses, I did not accept those and don't have to pay attention to
them -- that's because they're all license grants, not contracts to
which I have agreed.  They are disjunctive in combination, not
conjunctive.

> Compare this to the QPL permission to release a patch under any version you
> well please, as long as you don't do binary distribution.

That's uninteresting -- for Free software, we have to assume all users
modify the software and distribute modified versions to their close
friends.

-Brian

-- 
Brian Sniffen                                       bts@alum.mit.edu