GPLv3 comedy unfolding

[Alexander Terekhov]

http://gplv3.fsf.org/discussion-committees/B/Minutes/GPLv3Bconfcall_4_16mar2006/
http://gplv3.fsf.org/discussion-committees/D/logs/log20060321

Now, but this

http://gplv3.fsf.org/discussion-committees/D/logs/log20060228/

eposode is by far the best and deserves copy&paste.

Oh poor Java, PHP, and other "interpreters". ROFL.

Copyright © 2006 Free Software Foundation, Inc., 51 Franklin Street,
Fifth Floor, Boston, MA 02110-1301, USA. Verbatim copying and
distribution of this entire article are permitted worldwide, without
royalty, in any medium, provided this notice, and the copyright notice,
are preserved. 

-----
Log of 2006/02/28 meeting
by novalis — last modified 2006-03-07 15:45
22:04 <@novalis_dt> Let's start with a roll call. Who's here?
22:04 <@novalis_dt> David Turner, FSF
22:04 <+MarkDoliner> Mark Doliner, Gaim
22:04 <+fontana> Richard Fontana, SFLC
22:04 <+scalesda> Daniel Scales, Choate, Hall & Stewart
22:04 <+dondelelcaro> Don Armstrong
22:05 <+dondelelcaro> let me get the gobby going too
22:05 <@novalis_dt> thanks, Don.
22:05 <@novalis_dt> jblack, I know I saw you earlier...
22:05 <+jblack> James Blackwell, Bazaar-NG
22:06 <+dondelelcaro> ok... gobby is working
22:07 <@novalis_dt> OK. First order of business: any volunteers to
coordinate, or shall I?
22:07 <@novalis_dt> OK, any objections to me coordinating?
22:07 <+tomislav_medak> hi, guys
22:08 <@novalis_dt> Great, passed by consensus.
22:08 <+tomislav_medak> novalis_dt: no objection, it seems to be your
eternal doom
22:08 <@novalis_dt> Welcome to Masayuki Hatta, new member.
22:08 <@novalis_dt> Remind me of the name of the other guy we were
voting on?
22:09 <@novalis_dt> sedwards?
22:09 <+dondelelcaro> sounds right
22:09 <@novalis_dt> oh, nevermind, I found it.
22:09 <@novalis_dt> Unfortunately, I don't seem to be configured to view
results on that one.
22:09 <@novalis_dt> I'll play with it later.
22:10 <@novalis_dt> Anyway, jblack, you had taken a couple of issues. Do
you want to discuss them?
22:10 <+jblack> Not at this point. Its been a busy week for me, so I
didn't have time for the FAQ. :(
22:11 <+Rickerby> Hi Sorry I'm late
22:11 <@novalis_dt> Rickerby, No worries.
22:11 <@novalis_dt> Oh, we should also approve last week's minutes.
22:11 <@novalis_dt>
ttp://gplv3.fsf.org/discussion-committees/D/minutes/minutes_20060221
22:11 <@novalis_dt> Any comments on them before we approve them?
22:12 <@novalis_dt> Any objections to approving them?
22:12 <+dondelelcaro> they look fine to me
22:12 <+jblack> look good to me
22:12 <@novalis_dt> OK, if there are no objections, approved by
consensus.
22:13 <@novalis_dt> Don, any progress on the API key issue?
22:14 <@novalis_dt> One of the other committees also discussed a related
issue of authentication codes for cell phone networks, for instance?
22:15 <@novalis_dt> Don was going to look at the CCSC provision and try
to draw up a list of required and non-required keys, which we could then
build rules from.
22:16 <@novalis_dt> (is this thing on?)
22:16 <+MarkDoliner> Yes :-)
22:17 <@novalis_dt> OK, don is probably AFK.
22:17 <@novalis_dt> We'll move on and see what progress obra has made.
22:17 <@novalis_dt> Jesse?
22:18 <@novalis_dt> OK, is mako here?
22:18 <@novalis_dt> Ok, since nobody who has outstanding issues is
prepared, is there any new business? Any comments that people are
interested in?
22:19 <@novalis_dt> I have an item or two myself, which I haven't yet
added as comments.
22:19 <@novalis_dt> OK, here's my theory, and we'll see if anyone thinks
it's worth discussing:
22:20 * jblack listens
22:20 <@novalis_dt> I think the major components/system library
exception should permit all interpreter bindings.
22:20 <@novalis_dt> Let me explain.
22:20 <@novalis_dt> Let's say there's some GPL library -- libfoo.
22:20 <@novalis_dt> People want to access libfoo from Java, from PHP,
and from all manner of proprietary and GPL-incompatibly-licensed
languages.
22:21 <@novalis_dt> Even though the interpreters for these languages are
non-free, free software can still be written in the languages.
22:21 <@novalis_dt> I propose that any GPL library ought to be allowed
to link against one of these interpreters to the extent necessary to
expose the library functionality in the language.
22:22 <@novalis_dt> Now, this isn't a general-purpose loophole, because
it won't allow just any linking. Only bindings. And when code uses the
library through the bindings, it will still have to be GPL.
22:22 <@novalis_dt> So, I don't think this weakens copyleft any.
22:22 <+fontana> novalis_dt: How is this being treated under GPLv2?
22:22 <@novalis_dt> fontana, it's not permitted, because these
interpreters are incompatible with the GPL.
22:23 <@novalis_dt> fontana, Is my proposal clear?
22:23 <+jblack> My gutshot reaction would be no. I wouldn't want to help
make proprietary software better.
22:23 <+fontana> Yes, I understand it.
22:23 <+MarkDoliner> I think I kind of agree with jblack
22:23 <@novalis_dt> jblack: I'm not convinced that it helps proprietary
software very much at all.
22:24 <@novalis_dt> It makes it easier to write free programs in
languages without free interpreters.
22:24 <@novalis_dt> But it doesn't make it any easier to write non-free
programs.
22:24 <+MarkDoliner> It takes away the incentive for the interpreters to
be GPL compatible
22:25 <+MarkDoliner> But then I guess it promotes free software for
programs without free interpreters, so it seems like a trade off
22:25 <@novalis_dt> MarkDoliner, that's true -- but I think most people
who write libraries tend to allow linking with non-GPL-but-free
interpreters anyway.
22:25 <@novalis_dt> I don't know of anyone saying, for instance, "no,
you can't make PHP bindings for my GPL library."
22:26 <@novalis_dt> It's possible that it's happened and I didn't hear
about it.
22:26 <@novalis_dt> MarkDoliner, does GAIM have a library form?
22:26 <@novalis_dt> Or is it just a desktop app?
22:26 <+MarkDoliner> Not officially. We're working towards having a
"libgaim," and Adium uses a version of libgaim that they maintain
22:27 <@novalis_dt> And would it be your project's inclination to permit
PHP bindings for this? (for instance, for a web <-> AIM gateway).
22:28 <+jblack> Here's the problem with that. That proprietary product
is probably competing with equivilant free software, right? If you give
proprietary software the same access as the free software, they've just
lost an important edge.
22:28 <@novalis_dt> jblack, So, we're not giving people any incentive to
write in Python as opposed to PHP?
22:29 <@novalis_dt> jblack, That's a very good point.
22:29 <+MarkDoliner> I can't speak for everyone else, but I think most
developers would be in favor of allowing PHP bindings for it (although,
from a technical standpoint I don't think it would work at all because
of the way libgaim is architected, but that's not really important for
this discussion)
22:29 <@novalis_dt> MarkDoliner, What about Java bindings (assuming that
they only worked with Sun's Java)?
22:30 <+jblack> MarkDoliner: There's certainly a distiction between
incompatible software and nonfree software. PHP would probably be fine.
Sun's java though?
22:30 <+MarkDoliner> Ok, I guess I see your point
22:31 <+MarkDoliner> Linux is to an ELF executable as a Sun JVM is to
Java bytecode
22:32 <+MarkDoliner> And when you think of it that way, the Sun JVM
should be included in the OS library exclusion thingy
22:32 <@novalis_dt> MarkDoliner, the Sun java.lang.* is in there, too.
22:32 <@novalis_dt> MarkDoliner, the issue is not with pure-Java apps,
but with apps which link against the JVM.
22:33 <@novalis_dt> (or libraries that do so).
22:33 <@novalis_dt> Ok, I think I'm convinced that the general case is
not a great idea.
22:34 <@novalis_dt> If anyone isn't convinced, speak up now, or submit a
comment.
22:34 <+MarkDoliner> Heh heh
22:34 <@novalis_dt> But I do think that PHP license compatibility would
be nice.
22:34 <+jblack> Yeah! I do think there's room for stuff thats free.
22:35 <+jblack> Do we have an official definition of qualifications for
free software?
22:35 <@novalis_dt> In the current draft, a lot of licenses are
compatible under section 7.
22:35 <@novalis_dt> PHP isn't.
22:35 <@novalis_dt> jblack, Metalanguage for "free softare" isn't likely
to happen.
22:35 <+jblack> Then I'm convinced too.
22:36 <@novalis_dt> jblack, We've tried hard to figure out a way to say,
"this is compatible with all 'free software' licenses".
22:36 <@novalis_dt> MySQL tried too -- they ended up listing them.
22:36 <@novalis_dt> I've seen drafts of other licenses which tried to do
this, and they also ended up unsatisfying.
22:37 <@novalis_dt> I'm not saying it's impossible, but I don't think
anyone is likely to come up with it in the next year.
22:37 <+Rickerby> To me, the question becomes how do you allow
"coopition." Sometimes the best path to competing is to cooperate. How
does the free software community decide when they will benefit from
allowing development on a non-free platform with sufficient market
penetration? Or does the ideology demand purity?
22:37 <@novalis_dt> So, section 7 has been written to instead refer to
specific types of restrictions that we've seen in real licenses.
22:38 <@novalis_dt> Rickerby, it is an interesting question to ask
whether the Free Software community benefits from all of the Free Java
software which runs only on Sun's JVM.
22:38 <@novalis_dt> Rickerby, because GNU Classpath is always improving
22:38 <+fontana> novalis_dt: what is specifically the problem with the
PHP license for compatibility under GPLv3?
22:39 <@novalis_dt> And as it improves, software that was free but
"trapped" becomes truly free.
22:39 <@novalis_dt> fontana, 1 sec, checking...
22:40 <terjebr> What if I want to make GPLed software in C# and
microsoft .Net and link in some GPLed library?
22:41 <@novalis_dt> fontana, Section 4 of the php license in 3.0 seems
to go beyond trademark law.
22:41 <+fontana> aha.
22:41 <@novalis_dt> Motion to admit terjebr to this meeting?
22:41 <+jblack> second
22:41 <@novalis_dt> Any objections?
22:41 <@novalis_dt> OK, passed by consensus.
22:42 <terjebr> Thank you
22:42 <@novalis_dt> terjebr, BTW, can you introduce yourself?
22:42 <terjebr> I am Terje Bråten from Norway
22:42 <@novalis_dt> terjebr, if that library has to link against
Microsoft's interpreter, then no go.
22:42 <terjebr> I am much interested in programming free software on my
spare time
22:43 <terjebr> I know, but your suggestion was that we should open up
for that?
22:43 <@novalis_dt> terjebr, Yes.
22:43 <terjebr> I second that
22:44 <@novalis_dt> terjebr, Hm, OK. Would you like to submit it as a
comment through the comment system, or should I?
22:44 <terjebr> Then many people that program in .net and C# can
contribute to free programware and use GPLed libraries in their programs
22:44 <@novalis_dt> terjebr, Shouldn't they just choose a free language,
or a free interpreter for .net such as Mono?
22:45 <terjebr> Since it was your idea, I think you should submit it
22:45 <@novalis_dt> OK, I'll do that (noting that I no longer agree with
it)

22:45 <terjebr> Many programmers use C# because the work with it dayly
22:46 <@novalis_dt> soufron, welcome.
22:46 <terjebr> Their employer give them no choice in what environment
to use
22:46 <@novalis_dt> terjebr, Ok, so why not just use the Mono
interpreter, which is free?
22:47 <@novalis_dt> BTW, just got the results. Mike Nordell is approved.
22:47 <+mako> i'm not going to be able to participate
22:48 <@novalis_dt> mako, today, or at all?
22:48 <+mako> today
22:48 <terjebr> Ok, then you can only use the program with mono then
22:48 <@novalis_dt> mako, OK, we'll catch you next week.
22:48 <+mako> there's been a death in my research group
22:48 <@novalis_dt> terjebr, this is fine, right?
22:48 <@novalis_dt> mako, Sorry to hear about that.
22:48 <+mako> i should be around next week
22:48 <@novalis_dt> Hope it wasn't your advisor!
22:48 <+mako> yes.. we just found out
22:48 <+mako> no
22:48 <+mako> it wasn't
22:49 <+mako> talk to you all next week
22:49 <terjebr> I guess I have to think more about this
22:49 <@novalis_dt> terjebr, Ok. Please do submit comments in reply to
my comment (which I'll post after this is done).
22:51 <terjebr> I will post my comments, yes
22:51 <+jblack> I will as will
22:52 <@novalis_dt> fontana, the Zend Engine license also has an
original-BSD style advertising clause.
22:52 <terjebr> But I have no strong feelings about this, as opposed to
the DRM case
22:52 <@novalis_dt> terjebr, We invite your comments on DRM via the
comment system.
22:52 <terjebr> You have read my comment 902?
22:53 <@novalis_dt> I think the reason we chose not to have
compatibility with advertising clauses -- that is, those which control
the content of printed advertisements, is straightforward.
22:54 <+fontana> Yes.
22:54 <@novalis_dt> Has anyone noticed any comments to the contrary? Or
does anyone wish to make any which we can discuss as a committee?
22:56 <@novalis_dt> OK, so the response to the PHP issue seems to be
that nobody minds allowing linking with PHP (and Zend)-- but everyone is
skeptical of allowing linking with code with advertising clauses.
22:57 <@novalis_dt> MarkDoliner, jblack, how do you resolve this seeming
inconsistency?
22:57 <+jblack> I've never understood why the advertising clause is a
bad thing.
22:57 <+MarkDoliner> I don't know, I don't really mind NOT linking with
PHP, either :-)
22:58 <@novalis_dt> jblack, Imagine trying to write an advertisement for
the Linux kernel. "Good scalability (thanks to HP, IBM, J. Random
Hacker, X. Other. Hacker,...)
22:59 <+jblack> Yeah, I know there's a bit of a long list problem there.
But otherwise... its credit for work
22:59 <@novalis_dt> jblack, the problem isn't so much having to have one
bit of fine print at the end of your ad -- but that everyone who adds
some code gets to add one.
23:00 <@novalis_dt> jblack, Imagine a superbowl ad -- at 200,000 dollars
per second, those names get expensive fast.
23:00 <+jblack> Sure, but this is within the context of linking.
23:00 <@novalis_dt> jblack, let's say it's a SSL library.
23:00 <+jblack> so perhaps I mispoke. I know that there's problems with
ad clauses. But that doesn't make software evil.
23:00 <@novalis_dt> jblack, Oh, I agree.
23:00 <@novalis_dt> It's still Free Software.
23:01 <@novalis_dt> The question is, should we allow it to be linked to
GPL software?
23:01 <+dondelelcaro> it's really a lesser evil because those clauses
are annoying
23:01 <+dondelelcaro> the only issue where it's even interesting is
openssl, and gnutls has pretty much made that issue moot
23:02 <@novalis_dt> dondelelcaro, there's not a lot of software that
still uses the ad clause, true.
23:02 <+jblack> I suppose that the ad clause can be evil in some cases.
I suppose a hacker in china could go to jail for contributing to Tor,
for example.
23:02 <@novalis_dt> dondelelcaro, but you would pretty quickly get into
an advert envy situation where everyone would want to have them.
23:03 <+jblack> within the context of 3rd party linking thouh, it seems
neutral. I wouldn't use an ad clause as a reason for restriction.
23:03 <@novalis_dt> jblack, "third party"?
23:03 <@novalis_dt> jblack, Anyone who adds a module to the code gets to
add any restriction in section D.
23:04 <+jblack> For allowing linking with code.
23:04 <@novalis_dt> jblack, Huh?
23:04 <@novalis_dt> BTW, I have to take a call in 30 minutes. Just to
give a time limit here.
23:04 <+jblack> huh huh?
23:04 <+jblack> We're talking about allowing third party linking,
correct? or have we moved entirely to the Ad clause
23:05 <@novalis_dt> jblack, We're talking about section 7, which
doesn't, IIRC, say anything about third parties.
23:05 <@novalis_dt> In other words, you could be linking in some module
you downloaded off of Freshmeat, or you could be linking in one you
wrote yourself.
23:05 <+jblack> pardon, I'm off in the previous conversation
23:06 <terjebr> So, the problem is that you want a wider range of
various licensed interpreters to be able to link in GPLed libraries,
than the range og free interpreters that can be compatible with the GPL
23:06 <+jblack> I agree that we don't want advertising as a additional
restriction.
23:08 <@novalis_dt> terjebr has a point tha the contradiction between
not wanting general ad clauses, and wanting PHP (Zend), can be resolved
by noting that the ad clause there will only screw things up for the PHP
bindings -- not for the library as a whole.
23:08 <@novalis_dt> (if that makes any sense)
23:09 <@novalis_dt> The question, I suppose, is whether it's worth doing
a whole exception to section 7 just for PHP.
23:09 <+dondelelcaro> FE, the s/DEFANGED_//; change should probably not
have been made unless krooger had okayed it...
23:09 <+dondelelcaro> ECHAN.
23:10 <@novalis_dt> OK, does anyone have any more comments on PHP
compatibility?
23:10 <+dondelelcaro> not really; PHP has a few other problems beyond
the ad clause.
23:10 <@novalis_dt> dondelelcaro, Really?
23:11 <@novalis_dt> dondelelcaro, I mean, other than the ad clause and
the overbroad trademark bit.
23:11 <+dondelelcaro> novalis_dt: yeah, the overbroad trademark bit is
the one that I'm talking about...
23:11 <@novalis_dt> dondelelcaro, do you think that section 7 ought to
permit that?
23:11 <+dondelelcaro> my favorite example is that you can't call your
work "Telegraphpole" even though it clearly has nothing to do with
php...
23:12 <+dondelelcaro> or graphpad...
23:12 <+dondelelcaro> novalis_dt: no... I think that trademark law is
the proper domain for trademarks; legislating trademarks through
copyright is fundamentally flawed IMO.
23:13 <@novalis_dt> dondelelcaro, sure, it's a badly drafted license.
23:13 <@novalis_dt> But should we be compatible with it anyway given
that it's popular?
23:13 <+fontana> Are there any free software licenses that make explicit
exceptions in trademark restriction clauses for trademark fair use?
23:14 <@novalis_dt> fontana, one standard term in BSD-like licenses says
that you can't use the names of contributors to "endorse or promote"
23:14 <@novalis_dt> fontana, which gets pretty close.
23:14 <+fontana> OK.
23:14 <+dondelelcaro> dunno; I don't find the popularity argument
particularly convincing especially given that I've argued that overbroad
trademark clauses are non-free...
23:14 <@novalis_dt> dondelelcaro, I certainly could imagine that a
clause so overbroad as to be non-free.
23:15 <+dondelelcaro> novalis_dt: yeah, but that's illegal (or likely to
be illegal) in most jurisdictions anyway...
23:15 <+dondelelcaro> novalis_dt: right...
23:16 <@novalis_dt> dondelelcaro, You mean the BSD thing? Precisely,
which is why we think it's compatible with v2.
23:16 <+dondelelcaro> novalis_dt: yeah; that's been my argument for its
compatibility too.. so I think we're in agreement there.
23:17 <@novalis_dt> dondelelcaro, I'm not particularly bothered by
"telegraphpole".
23:17 <+dondelelcaro> novalis_dt: well, it definetly poses a problem for
the DFSG freeness of a work; which is really the basis which I'm working
from.
23:17 * MarkDoliner runs off to another engagement
23:17 <@novalis_dt> dondelelcaro, that depends on your reading of DFSG.
:)
23:17 <+dondelelcaro> novalis_dt: sure...
23:18 <terjebr> What is DFSG?
23:18 <+dondelelcaro> terjebr: debian free software guidelines
23:18 <terjebr> Ok, thanks
23:19 <@novalis_dt> Does anyone want to do something about the PHP
thing?
23:19 <@novalis_dt> Or should we simply note in minutes that we
discussed it and decided to leave it as-is?
23:19 <+dondelelcaro> I should note too that there are qa few cases
where I think things are non-free which comply with the DFSG... just so
no one is surprised when you see me on -legal saying something is DFSG
free when I've complained about it's non-freeness. ;-)
23:20 <@novalis_dt> qa?
23:20 <soufron> hello
23:20 <@novalis_dt> hello, soufron
23:20 <+dondelelcaro> novalis_dt: started out as quite a few, and I
deleted most of the quite, but not all of it. ;-)
23:20 <soufron> I wanted to ask for someone else to be elected to the
committee
23:20 <soufron> but maybe that's not the right time
23:21 <@novalis_dt> soufron, Let's go ahead and just do the last few
additions right now.
23:21 <@novalis_dt> soufron, you have one, and I have one.
23:21 <@novalis_dt> soufron, Who is it?
23:22 <@novalis_dt> I propose Rowan Wilson.
23:22 <soufron> it's roberto di cosmo
23:23 <soufron> he's a professor of computer science at the university
of paris-VII
23:23 <soufron> and he is the director of the open source master
23:23 <@novalis_dt> soufron, OK, sounds good to me.
23:23 <soufron> even if I don't know how to call it
23:23 <soufron> basically I teach open source law in his master
23:23 <soufron> and he's good
23:23 <soufron> just google him
23:23 <@novalis_dt> Any objections to just doing a quick poll here on di
Cosmo?
23:23 <@novalis_dt> Or do we want to go through the whole process?
23:23 <+fontana> I myself have no objections, but can we consult with
Eben Moglen before officially adding new committee members?
23:24 <@novalis_dt> fontana, Sure.
23:24 <@novalis_dt> Rowan Wilson is with a UK Free Software group, and
he has been interested in attending for a while.
23:24 <soufron> still, I have to go now...
23:24 <soufron> I'll be back in a few 20 minutes
23:24 <soufron> ok ?
23:24 <@novalis_dt> soufron, We will be over then.
23:24 <@novalis_dt> soufron, So, see you next week.
23:25 <soufron> maybe we could vote on roberto before I go...
23:25 <soufron> he's a famous french free software proponent
23:25 <@novalis_dt> soufron, fontana wishes to discuss it with Eben
first.
23:25 <terjebr> Beware that this irc log will be public, and it may not
be a good idea to discuss such things in public
23:25 <soufron> I will see Eben next week
23:25 <@novalis_dt> terjebr, We know. There is a procedure.
23:25 <soufron> but anyway
23:25 <@novalis_dt> soufron, So, let's put this on hold until we have
that coversation.
23:25 <soufron> roberto cannot participate right now anyway
23:26 <soufron> no problem
23:26 <soufron> just keep his name in your minds
23:26 <soufron> ok ?
23:26 <@novalis_dt> Sounds good.
23:26 <soufron> see you later all
23:26 <@novalis_dt> My time is running out, so does anyone have any
comments before we adjourn?
23:26 <@novalis_dt> next week, we look forward to hearing from a subset
of [mako, jblack, don, obra].
23:26 <+jblack> No I.
23:27 <@novalis_dt> next week, we look forward to hearing from a subset
of [mako, jblack, don, obra].
23:27 <@novalis_dt> er
23:27 <@novalis_dt> OK, adjourned.
-----

regards,
alexander.