Re: LGPL vs. GPL

[JohnF]

Ciaran O'Riordan <ciaran@fsfe.org> wrote:
> Hi John,
> These cases are never black and white, and I don't know PocketCAS or
> MimeTex, so I can't give any advice on this situation, but here are some
> general ramblings anyway...

Hi Ciaran,
     Sure.  Some general ramblings are exactly what I was looking for.
And thanks again for your time and effort.
     Fyi, as far as I can tell, what this guy (Daniel Alm) appears to
be thinking about is writing some glue code that ties together the
xcas/giac packages he cited with mimetex.  And I think that should
be pretty trivial after he's figured out where to best place all the
hooks he'll need.  But figuring that out might or might not be a pain
(just piping one program's output to another's input should be trivial).
     I also can't tell whether Alm is serious about actually doing
something, or whether he's just talking.  My (limited) experience
has been that serious people contact me after they've already written
at least something.  But as far as I can tell, the only thing he's
done so far is email me.

> If PocketCAS is written to specifically work with MimeTex, then PocketCAS
> might be a "derived work" which would mean he needs your permission to
> distribute PocketCAS.  Because your software is GPL'd, "needing your
> permission" means he can either (a) distributing his software under the GPL
> or a GPL compatible licence such as the LGPL or Revised BSD or (b) ask you
> for an exception.
> 
> If PocketCAS only performs simple data exchange with MimeTex, such that
> other applications could be substituted for MimeTex, then it's likely that
> PocketCAS is not a derived work and so it does not need a GPL compatible
> licence.

I'm aware of these kinds of ambiguities (whether or not a specific
instance of using gpl'ed code remains compliant with the gpl), and
not at all competent to resolve them myself.  My usual attitude 
has been to answer such requests as though my specific permission 
is actually needed (as though the intended use is not gpl compliant).

>> I'm inclined to say, "Go ahead and distribute a binary image of mimetex
>> along with your PocketCAS."  Is there any open-source-related reason 
>> (or any other reason) I shouldn't say that?
> 
> One thing that's for sure is that he'll have to distribute MimeTex's source
> with the binary (or an offer to send people the source on request).

I'd thought a link to its homepage (where the source can be downloaded)
satisfies that requirement.

> About his request for clarification of which version of GPL MimeTex is
> under.  It's a good idea to answer that request anyway.  Version 3 is the
> current version, and here's an explanation of the improvements in it:
> http://www.gnu.org/licenses/quick-guide-gplv3.html

The earliest gpl'ed version was released in 2002 under gplv2.
The last significant change was in 2006, still under gplv2.
If a non-trivial opportunity presents itself, I'll release
it under gplv3.

>> Should I maybe instead
>> say something like, "Permission is granted to distribute mimetex with
>> free versions of PocketCAS, but not with commercial versions."?
> 
> That probably isn't a good idea.
> 
> You could only make such a demand *if* he does need your permission to
> distribute PocketCAS (i.e. if PocketCAS is a derived work of MimeTex).  And
> if that was the case, it would be better to inform him that he has to
> distribute his software under the GPL.  Granting exceptions lessens the
> incentive for any related project to release their software as free
> software.
> 
> As for the general request, if this developer is not helping you in any way,
> and since he's not helping society in any way, I don't see the incentive to
> give him the explicit permission he's asking for.

The free-of-charge version of his program helps society (assuming it's
useful in the first place), though it doesn't help me (beyond just using
it for free if it's useful).  At least, that's what I was thinking.
If that's wrong, then you're right that I have zero incentive.
     Now, I'd guess his principal incentive is the not-free-of-charge,
commercial version, that I'm not granting permission for.  So maybe
he now has zero incentive, too.  On the other hand, he might just go
ahead with the free version, and maybe eventually even gpl it.

> Getting a definitive answer would take time/work and might need a lawyer.
> (And you would need a copy of PocketCAS so that you could see how it
> interacts with MimeTex.)  And since he's the benefactor, this work/cost
> should be his.

Yeah, that's pretty much my major gpl gripe.  People keep emailing me
whether such-and-such a use is permitted.  My typical, and most recent
answer takes the form (excerpted from a recent email)...

     The "not quite clear-cut" (your words) aspects of your plans,
  which you describe below, are your responsibility to determine.
  What you're asking me to do is to legally interpret the gpl for you.
  But I'm not an expert on that.  Perhaps you think it's strange
  that someone (me) would release a program under a license they
  don't completely understand.  But that's not uncommon with the gpl,
  where people have trusted the fsf to devise a license that
  reflects the spirit of the open source community.
     So, as noted above, I've cc'ed the gpl compliance experts for you.
  I'll leave it to you and them, and any other gpl experts you care to
  consult, what you must do to make your commercial programs gpl compliant.
  If you choose to do that, then you're licensed to incorporate mimetex
  (and, for that matter, any other gpl'ed code).  If you choose not to do
  that, then you're not legally licensed to incorporate any gpl'ed code.

That's one big pain.  I wish the fsf would think about some effective
way for gpl authors to avoid these kinds of emails.  As noted, I just
cc compliance@fsf.org, and hope the emailer will go away.  It's not
an issue I want to deal with or am competent to deal with.  But using
the gpl invites exactly these kinds of issues.

> So in general I'd recommend against granting such exceptions, but without
> the details, I can't give real advice, and it's your decision anyway.  At
> the very least, I'd ask that every place that says "Copyright PocketCAS",
> should also say "MimeTex is copyright John Forkosh can be redistributed as
> free software under the GNU GPL v3" (assuming v3 is the version you're
> using).  This is to avoid the situation where users either don't know that
> it's MimeTex (not PocketCAS) that is producing the high quality output, or
> without knowing that they can get the source, modify, contribute, and
> redistribute MimeTex.

Sounds good.  And thanks yet again, Ciaran, for your time and effort,
-- 
John Forkosh  ( mailto:  j@f.com  where j=john and f=forkosh )