rms@AI.MIT.EDU (06/08/89)
Ok, now I have a question. What would be the attitude toward a developer who distributes .o files containing proprietary code, source to the GNU libraries in question, and leaves the end user to compile and link them together on his own? If these libraries are standard, then it would be hard to argue that the developer is doing anything which intrinsically relates to the GNU libraries in question. He might not even know whether users choose to link with GNU libraries or other libraries. So this must be permitted. If the libraries involved were customized out of GNU software by the developer for this one application, then there would be an argument that they really are parts of one large program. My lawyer says it isn't clear what the law says about this. Since aiding hoarders is not my aim, I don't see a point in conceding this particular issue in advance.
grunwald@flute.cs.uiuc.edu (Dirk Grunwald) (06/08/89)
In article <8906071734.AA00295@sugar-bombs.ai.mit.edu> rms@AI.MIT.EDU writes:
Ok, now I have a question. What would be the attitude toward a developer
who distributes .o files containing proprietary code, source to the GNU
libraries in question, and leaves the end user to compile and link them
together on his own?
.......
My lawyer says it isn't clear what the law says about this. Since
aiding hoarders is not my aim, I don't see a point in conceding this
particular issue in advance.
-----
The issue is, in fact, important, if not to the software horders, than to
the people who develop Gnu library routines.
Say I've put some code in libg++. By putting in the ``public trust''
of FSF, I'm hoping that it will benefit all. If anyone is going to
make money off of it by hording software, I'd rather it was me.
Unless those library copylefts are bullet-proof, the copyleft provides
the illusion of protection without the mechanism of enforcement. If they
are not bullet-proof, they should either be changed or the attitude to
libraries should be changed.
Perhaps the libraries should be licensed products, with a specific
re-distribution clause that precludes these sorts of activities.
Does your control increase when you license software, rather than
copyright it?
The hard part is making it easy enough to get the license if you're
``a good guy'' but still having enough teeth to bite ``a bad guy''.
--
Dirk Grunwald -- Univ. of Illinois (grunwald@flute.cs.uiuc.edu)
kevin%kalli@Sun.COM (Kevin Sheehan {Consulting Poster Child}) (06/08/89)
In article <8906071734.AA00295@sugar-bombs.ai.mit.edu> rms@AI.MIT.EDU writes: > > Ok, now I have a question. What would be the attitude toward a developer > who distributes .o files containing proprietary code, source to the GNU > libraries in question, and leaves the end user to compile and link them > together on his own? > >If these libraries are standard, then it would be hard to argue that >the developer is doing anything which intrinsically relates to the GNU >libraries in question. He might not even know whether users choose to >link with GNU libraries or other libraries. So this must be >permitted. > >If the libraries involved were customized out of GNU software by the >developer for this one application, then there would be an argument >that they really are parts of one large program. > >My lawyer says it isn't clear what the law says about this. Since >aiding hoarders is not my aim, I don't see a point in conceding this >particular issue in advance. I agree that hoarding is Bad, but I'm compelled to point out that the rest of the world doesn't have the freedom to consult for a living, and do Great Software on the side. Purely hardware companies (the orignal target of your actions in some sense) have historically kept things proprietary to keep people buying their hardware, and keep them paying Big Bucks for support and availability. A legitimate target for free software, and if the world had been different years ago, probably a beneficiary considering the redundant development efforts done over the years. But they were locked into competing with Big Blue on its terms and got stuck- I sincerely hope you can win that battle. However, what about all the companies that put honest sweat and effort into their product? Folks that fund development with nominal charges? They are all restricted from using the FSF software - which is where I disagree with the copyleft - I disagree with usury, but forcing folks who use the tools to create their product to follow along with free is a bit hampering IMHO. Imagine if using the worlds greatest screwdriver obligated you to give away the system it built - if using a 68K chip obligated you to give it and the schematics out. Hitting companies by issuing better quality software for free as a method of forcing them to adhere to more polite modes of business is useful for software dependent iron companies and greedy software companies. FSF software can only help Apple, an iron company. But, IMHO FSF software shouldn't penalize companies trying to use it in an honest day's work. If it's derived, you're entitled to demand your due from copyright law. If it's used as a tool, I really think you ought to let it go. In my case, you'd have five more companies completely sold on the idea, if their lawyer hadn't pissed his pants first... Kevin Sheehan Sun Microsystems
kevin%kalli@Sun.COM (Kevin Sheehan {Consulting Poster Child}) (06/08/89)
In article <GRUNWALD.89Jun7135053@flute.cs.uiuc.edu> grunwald@flute.cs.uiuc.edu writes: > >Perhaps the libraries should be licensed products, with a specific >re-distribution clause that precludes these sorts of activities. >Does your control increase when you license software, rather than >copyright it? > >The hard part is making it easy enough to get the license if you're >``a good guy'' but still having enough teeth to bite ``a bad guy''. Yep - that's what Sun did with NFS and RPC. NFS was licensed bigtime, RPC is "freely licensed", which isn't the same as PD. Kevin Sheehan Sun Microsystems PS I second a vote for gnu.politics... I *love* g*, but this is getting a bit thick.
rms@AI.MIT.EDU (06/08/89)
Imagine if using the worlds greatest screwdriver obligated you to give away the system it built This analogy is false and misleading. USING most GNU tools imposes no restrictions. Only copying them or parts of them does so. In a couple of cases, bison and the libraries, the way to use them involves copying parts of them. In most cases, including gcc which is the subject for this list, it does not. - if using a 68K chip obligated you to give it and the schematics out. This analogy is better. It is accurate to a limited extent. In a large system, it is perfectly possible for GNU programs to coexist with separate proprietary programs. It is only within a single program that this analogy holds. IMHO FSF software shouldn't penalize companies trying to use it in an honest day's work. GNU software does not penalize anyone. It does, however, help some projects more than others. It discriminates against hoarders. GNU software can be used in certain ways even by hoarders. In certain other ways, it can be used only by sharers. This is what I intended. I have explained my reasons for this elsewhere, so I won't repeat them here. Many other contributors to GNU, who do not share my opposition to proprietary software, have expressed the opinion that, if something they wrote is going to be used in a proprietary program, they ought to get paid. They are willing to share with sharers like me, but feel no obligation to contribute to a proprietary program gratis. We all know that some people (many people) disagree with me. Some of them have already said so. If you are another one of these people, please spare the readers another message to this effect. Your opinion has already been expressed--saying it again won't help anyone.
pcg@aber-cs.UUCP (Piercarlo Grandi) (06/09/89)
In article <8906080326.AA00639@sugar-bombs.ai.mit.edu> rms@AI.MIT.EDU writes:
Imagine if using the worlds greatest screwdriver obligated you to give
away the system it built
This analogy is false and misleading. USING most GNU tools imposes no
restrictions. Only copying them or parts of them does so. In a
couple of cases, bison and the libraries, the way to use them involves
copying parts of them. In most cases, including gcc which is the
subject for this list, it does not.
Let me support RMS on this. If "hoarders" want to use FSF sw, without taking
undue advantage of it (selling it and restricting its redistribution), they
can. The *only* problem I can see is the use of FSF libraries or components
(whether in source form, e.g. bison, or compiled, e.g. libg) incorporated
in proprietary products; this is not impossible to solve, you can always use
PD libraries, or rewrite them, or buy equivalent ones (and if you'd rather
not, reflect that as a rule you cannot have your cake *and* eat it -- free
sw and sell it -- even if there are exceptions like X :->).
Finally, a proof of existence for those that still have doubts: NeXT.
Enough said?????????
--
Piercarlo "Peter" Grandi | ARPA: pcg%cs.aber.ac.uk@nsfnet-relay.ac.uk
Dept of CS, UCW Aberystwyth | UUCP: ...!mcvax!ukc!aber-cs!pcg
Penglais, Aberystwyth SY23 3BZ, UK | INET: pcg@cs.aber.ac.uk
brooks@vette.llnl.gov (Eugene Brooks) (06/11/89)
In article <999@aber-cs.UUCP> pcg@cs.aber.ac.uk (Piercarlo Grandi) writes: > >Finally, a proof of existence for those that still have doubts: NeXT. When NeXT distributes their NeXT Step code as copylefted source I will eat my hat! I have also heard that NeXT is exploiting the user does the link trick, sounds like software hoarder to me. Is RMS going to boycott them for this??? brooks@maddog.llnl.gov, brooks@maddog.uucp
honey@mailrus.cc.umich.edu (peter honeyman) (06/11/89)
here's what happens when you run the c compiler after installing vers 0.9: The current licensing agreement with the Free Software Foundation prevents NeXT from completely installing the C compiler. We hope to resolve this before the 1.0 release. To complete the installation of the C compiler, please run the command: installcc i guess NeXT has lawyers too. peter
pcg@aber-cs.UUCP (Piercarlo Grandi) (06/13/89)
In article <26713@lll-winken.LLNL.GOV> brooks@maddog.llnl.gov (Eugene Brooks) writes: In article <999@aber-cs.UUCP> pcg@cs.aber.ac.uk (Piercarlo Grandi) writes: > >Finally, a proof of existence for those that still have doubts: NeXT. When NeXT distributes their NeXT Step code as copylefted source I will eat my hat! Context: All this discussion started because some (greedy) people pointed out that they were prevented by the GPL *conditions* (none of them is a *restriction*, let me insist) from selling code compiled with FSF sw. RMS has many times pointed out that this is not true; only if they want to incorporate FSF code (e.g. the non essential libraries, the essential ones have been put in the PD by the FSF) in their products the GPL applies. Discussion: NeXT is living proof that you can *use* FSF sw to develop proprietary products. NeXT Step code was developed with gcc, gdb, gemacs, etc..., and has been sold for $10,000,000 to IBM after all. I have also heard that NeXT is exploiting the user does the link trick, sounds like software hoarder to me. Inasmuch their sw is not freely redistributable in source code, they are "hoarders" by definition, whether they use proprietary or free sw to develop it. The link trick is legitimate, after all. The GPL condition that any sw incorporating free sw *and distributed to third parties* must be free as well is designed, as RMS has said many times, only to *encourage* users of free sw to make their sw free as well, by making it more trouble to distribute proprietary sw using free one. Is RMS going to boycott them for this??? As far as I understand him, RMS is disappointed when "hoarders" take advantage of free sw to develop proprietary sw, but consistently with his principles he does not want to restrict free sw only to "sharers". Also, he has repeatedly said that the FSF are too busy/poor to get annoyed at all those that do something they don't like; the perpetrators of potential industry wide disasters like Apple are another matter though. -- Piercarlo "Peter" Grandi | ARPA: pcg%cs.aber.ac.uk@nsfnet-relay.ac.uk Dept of CS, UCW Aberystwyth | UUCP: ...!mcvax!ukc!aber-cs!pcg Penglais, Aberystwyth SY23 3BZ, UK | INET: pcg@cs.aber.ac.uk
gjc@bu-cs.BU.EDU (George J. Carrette) (06/13/89)
If an operating system implements sharable libraries, like VMS has done for years, and like SUN is doing (modulo bugs) in their release 4.0 then "the user does the link" legal problem does not come up. Even the lowly AMIGA has a sharable library construct. Perhaps NeXT can manage to implement one too. You really need it badly when you have large libraries, e.g. Xwindows. -gjc
benson@odi.com (Benson Margulies) (06/14/89)
In article <32893@bu-cs.BU.EDU> gjc@bu-cs.BU.EDU (George J. Carrette) writes: >If an operating system implements sharable libraries, like VMS >has done for years, and like SUN is doing (modulo bugs) in >their release 4.0 then "the user does the link" legal problem >does not come up. Funny thing. In private communication, RMS said that just the opposite was true, at least for now. I SURE WISH THAT UNIMFORMED FSF ZEALOTS WOULD RESIST THE URGE TO POST. Mr. Tower nearly announced the imminent publication of "chairman stallman's little red book," in his adualatory posting about RMS. Well, its clear that as much bullshit has been posted in this discussion by second-order FSF zealots as by FSF non-believers. It also seems to me that some of the doubters are at least as intelligent as literate as RMS. If the FSF wants to use the news mechanism to distribute a controlled discussion, you-all should make the groups moderated. I volunteer to send out the control messages if you don't know how. If you don't want to moderate it, then you are playing King Canute in asking news readers to read carefully, respond moderately, or learn to spell. And while I'm annoyed: when is the FSF going to remove the anti-semitic illustration from the cover of the emacs manual? The supposed software hoarder could be copied from any number of Nazi cartoons in the museum of them in Jerusalem. -- Benson I. Margulies
gjc@bu-cs.BU.EDU (George J. Carrette) (06/14/89)
I am no FSF zealot, and I stand by my observation that sharable libraries remove the problem of encorporating copyrighted library code, no matter what your interpretation of a private RMS communication happens to be. Put up or shut up. What did RMS say and how do you interpret it to prove that what I said is, to quote your foul mouth, "bullshit"? Also, your reference to the biblical zealots is way off base. It would be more accurate to compare FSF followers with the followers of new testament figures. -gjc
pardo@june.cs.washington.edu (David Keppel) (07/30/89)
rms@wheaties.ai.mit.edu writes: >The FSF lawyer told me that distributing code which was designed to >work only when linked with part of a GNU program would be considered, >legally, as a subterfuge for distributing a modified program >containing both that and the GNU program. So it would be covered by >the general public license. This sounds like a `proprietary interface' argument. ``You must follow the copyleft on all code that uses something other than the standard (documented external) interface.'' Is this true? Followups to gnu.misc.discuss. ;-D on ( And if the program does the link... ) Pardo -- pardo@cs.washington.edu {rutgers,cornell,ucsd,ubc-cs,tektronix}!uw-beaver!june!pardo
grunwald@flute.cs.uiuc.edu (Dirk Grunwald) (07/30/89)
This does not protect the Gnu C library then; only the Gnu G++ library. If the Gnu C library is POSIX/ANSI/WHATEVER compatible, the distributor could claim that *any* library could be linked. For libg++, the interface is more detailed; it would be demonstrable that the user was intended to link the Gnu library, as opposed to, e.g., the AT&T library, into a product. -- Dirk Grunwald -- Univ. of Illinois (grunwald@flute.cs.uiuc.edu)