Zuul? Why not a nice Irish name like Patsy?) (09/11/89)
I just read in the EE Times (last week's issue, p. 108) that there's a group called the Free Software Foundation in Cambridge which has a C++ compiler/debugger. This stimulated a couple of questions: 1. Are there any users on the net who have used their compiler/debugger? 2. Is it really free, or is that just their title? 3. Is your personal experience with a SUN? If not, then what sort of system? 4. If you've used their stuff, would you recommend it? Thanks! Jim Kohli """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""" "What a waste it is to lose one's mind--or not to have a mind. How true that is." - V.P. Dan Quayle (possibly from self-experience) """""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""" Jim Kohli | "Stream? More like a drying trickle GE Medical Systems | of consciousness, if you ask me..." PO Box 414 | Milwaukee, WI 53201-414 | ------------------------------------------------------------------------- gemed!hal!kohli@crd.ge.com sun!sunbird!gemed!hal!kohli -------------------------------------------------------------------------
sarrett@ics.uci.edu (Wendy Sarrett) (09/11/89)
I have just started using the "Free Software Foundation" compliler called g++. It is reasonably good on the suns but we're waiting for the new release for our sequents. It has lots of library classes which are useful. Furthermore, it is a real c++ compiler as opposed to CC which translates c++ to c and then compiles the c. Good luck, Wendy
jima@hplsla.HP.COM (Jim Adcock) (09/12/89)
>1. Are there any users on the net who have > used their compiler/debugger? I have used g++ and gcc, and others here swear by their debugger. G++ generates pretty good code, but slightly more efficient code is generated by using AT&Ts cfront compiler as the front end and gcc as the back end. >2. Is it really free, or is that just their title? Depends on how you want to interpret "free." FSF requires no fee for use of their compilers, but do restrict what you can do with their compiler, and also restrict what you can do with code compiled by their compiler. There are workarounds to these restrictions, but the restrictions tend to be a problem for commercial users. Also FSF info speaks of a moral obligation to support them if you use their products. >4. If you've used their stuff, would you recommend it? Not for commercial work. For acedemic work, or playing around, maybe. Note, the AT&T compiler tends to be a little ahead of g++ in terms of having the latest language features implemented. >Jim Kohli ------ >Furthermore, it is a real c++ compiler as opposed >to CC which translates c++ to c and then compiles the c. > >Wendy CC is a real compiler that is targeted to a "C-language-machine." It uses "C" as the generated assembly language, if you will. I believe some companies have already successfully retargeted the AT&T compiler to generate proprietary assembly language for their particular CPUs. The major advantage of doing this is to allow vendor-specific debugging tools to be used. Slight performance gains may also be possible.
kohli@gemed (Jim Kohli, but my friends call me) (09/13/89)
I would like to thank you all for the informative and gentle-natured replies I have received regarding the Free Software Foundation. I would also like to apologize for asking what must have, in retro- spect, seemed like (was) an incredibly ignorant question. Again, thanks! Jim Kohli P.S. If there's anyone else out there that would like to know what this is all about, feel free to send me an inquiry and I'll try to send you my summary of responses (avoiding all personal references which wouldn't be common knowledge).
tvf@cci632.UUCP (Tom Frauenhofer) (09/19/89)
In article <6590249@hplsla.HP.COM> jima@hplsla.HP.COM (Jim Adcock) writes (in reply to another's questions) > >>1. Are there any users on the net who have >> used their compiler/debugger? >I have used g++ and gcc, and others here swear by their debugger. G++ generates >pretty good code, but slightly more efficient code is generated by using AT&Ts >cfront compiler as the front end and gcc as the back end. >>4. If you've used their stuff, would you recommend it? > >Not for commercial work. For acedemic work, or playing around, maybe. CFRONT 1.2 was DOA at many sites due to its many bugs. I don't know about CFRONT 2.0, but I have a general distrust for any release of software that is numbered X.0 (from my experience, first major releases of a software package are fairly buggy. I'm not claiming CFRONT 2.0 is buggy, just stating my experience). I've used g++ and found it (and most other GNU packages) to be solid tools. And the bug-fix turn-around time is quite short. And you get source. And the price is reasonable. >Note, the AT&T compiler tends to be a little ahead of g++ in terms of having >the latest language features implemented. No argument here, but the lag time between AT&T adding new language features and their incorporation into g++ is not all that great. >>Furthermore, it is a real c++ compiler as opposed >>to CC which translates c++ to c and then compiles the c. >>Wendy >CC is a real compiler that is targeted to a "C-language-machine." It uses >"C" as the generated assembly language, if you will. CC invokes the CFRONT pre-processor that converts your c++ code into something that can be digested by the C compiler. I can accept this definition of compiler providing you now consider CPP (the C pre-processor) is a compiler as well. It uses "C" as the generated assembly language, too. On the other hand, g++ generates {assembly? Object? I forget} for the target system. I find it easier to debug using g++/gdb than using CC/whatever-debugger-you-have-on-hand. >I believe some companies >have already successfully retargeted the AT&T compiler to generate >proprietary assembly language for their particular CPUs. The major advantage >of doing this is to allow vendor-specific debugging tools to be used. >Slight performance gains may also be possible. This could be a win, depending on the implementation. Unfortunately, you have to be using one of these hardware platforms. Which vendors are doing this? I am curious. I'm not flaming or cussing or the like, just playing devil's advocate. I just feel that the GNU tools got a shorter shrift than they deserve. Thomas V. Frauenhofer ...!rutgers!rochester!cci632!ccird7!tvf *or* ...!rochester!kodak!swamps!frau!tvf "The Earth? I'm going to blow it up. It obstructs my view of Venus" - Martin
jima@hplsla.HP.COM (Jim Adcock) (09/21/89)
>>Not for commercial work. For acedemic work, or playing around, maybe. > >CFRONT 1.2 was DOA at many sites due to its many bugs. I don't know about >CFRONT 2.0, but I have a general distrust for any release of software that >is numbered X.0 (from my experience, first major releases of a software >package are fairly buggy. I'm not claiming CFRONT 2.0 is buggy, just >stating my experience). In my experience, I have found 2.0 to be much more reliable than the prior AT&T releases, and more reliable than g++... ...Which is not why I don't recommend g++ for commercial work. I don't recommend g++ for commercial work because then your company's copyright gets entangled with the Gnu copyleft restrictions. I'm sure a sufficiently clever lawyer could get around this, but what would be the point? You'd be better off (as a commercial user) to pay AT&T and be free to use the generated code as you please. >CC invokes the CFRONT pre-processor that converts your c++ code into >something that can be digested by the C compiler. I can accept this >definition of compiler providing you now consider CPP (the C pre-processor) >is a compiler as well. It uses "C" as the generated assembly language, too. >On the other hand, g++ generates {assembly? Object? I forget} for the target >system. I find it easier to debug using g++/gdb than using >CC/whatever-debugger-you-have-on-hand. CPP is a macro processor. You can see this by making errors in your C code that passes silently through CPP, only to cause your C compiler to barf. Cfront never generates C code that causes your C compiler to barf -- except in the rare cases where there is either a compiler bug in cfront, or a compiler bug in your C compiler. Likewise a standard C compiler generates assembly that your assembler turns into machine code -- if your C compiler ever generated assembly that your assembler couldn't handle -- then that would represent a bug in either your C compiler, or your assembler. True compilers are distinguished by always generating correct output, or nothing at all [barring compiler bugs :-]
tvf@cci632.UUCP (Tom Frauenhofer) (09/22/89)
In article <6590261@hplsla.HP.COM. jima@hplsla.HP.COM (Jim Adcock) writes: >In a previous article, I wrote .>.Not for commercial work. For acedemic work, or playing around, maybe. .>CFRONT 1.2 was DOA at many sites due to its many bugs. I don't know about .>CFRONT 2.0, but I have a general distrust for any release of software that .>is numbered X.0 (from my experience, first major releases of a software .>package are fairly buggy. I'm not claiming CFRONT 2.0 is buggy, just .>stating my experience). .In my experience, I have found 2.0 to be much more reliable than the .prior AT&T releases, and more reliable than g++. Fine. That is good. How about more pro/con from the rest? P.S.: A representative from AT&T has been taking me to task about my statement about X.0 releases and that since I have no experience with CFRONT 2.0, then I should be quiet about the matter. If most others have the same good things to say about CFRONT 2.0, then AT&T has improved their software quality assurance, and other companies could take some notes from them (when it comes to major releases of software, at least). ....Which is not why I don't recommend g++ for commercial work. I don't .recommend g++ for commercial work because then your company's copyright .gets entangled with the Gnu copyleft restrictions. I'm sure a sufficiently .clever lawyer could get around this, but what would be the point? You'd .be better off (as a commercial user) to pay AT&T and be free to use .the generated code as you please. Only using the libg++ library does the GNU copyleft problem arise. .>CC invokes the CFRONT pre-processor that converts your c++ code into .>something that can be digested by the C compiler. I can accept this .>definition of compiler providing you now consider CPP (the C pre-processor) .>is a compiler as well. .CPP is a macro processor. You can see this by making errors in .your C code that passes silently through CPP, only to cause your C .compiler to barf. Cfront never generates C code that causes your C .compiler to barf -- except in the rare cases where there is either a .compiler bug in cfront, or a compiler bug in your C compiler. Likewise I will take back publicly (to the happiness of the AT&T person who has been mailing me his comments) that CPP is a compiler if CFRONT is. .True compilers are distinguished .by always generating correct output, or nothing at all [barring compiler .bugs :-] Yes, but unfortunately, enough untrue products from various vendors have come onto the market to make cynics of us all (please take note of this, my friend at AT&T). Thomas V. Frauenhofer ...!rutgers!rochester!cci632!ccird7!tvf *or* ...!rochester!kodak!swamps!frau!tvf "The Earth? I'm going to blow it up. It obstructs my view of Venus" - Martin
jima@hplsla.HP.COM (Jim Adcock) (09/27/89)
>Only using the libg++ library does the GNU copyleft problem arise.
I dispute this. [but note I'm not a lawyer so what I'm really saying is
go ask your lawyer which bring me back to my original statement that using
g++ on commercial work entangles you in copyright issues so why not just
use AT&T's [or other commercial] compiler.]
I believe _main and builtin_new, and .h files at the very least, also
complicate the issue. Even the templates by which C++ code get converted
to a particular CPU's assembly language could cloud these issues.
Stallman's statements not withstanding, I do not believe issue of copyright
infringement where a compiler "automatically" injects a greater or lessor
amount of its own source code into a user's work to be compiled/assembled
has been legally tested. "Fair Usage" rules certainly apply here, but what is
"fair use" when an author publicly states he doesn't really want his
work used for commercial purposes? ...I suspect the courts might find a very
narrow view of "fair use" in those circumstances.
..........In any case, if you want to use g++ for commercial work only a
competent patent/intellectual property lawyer can decide for you and your
company. Just make darned sure you and the lawyer know the intimate details
of how g++, .h files, porting to multiple machines, etc, work.mike@thor.acc.stolaf.edu (Mike Haertel) (09/28/89)
In article <6590268@hplsla.HP.COM> jima@hplsla.HP.COM (Jim Adcock) writes: >>Only using the libg++ library does the GNU copyleft problem arise. >I dispute this. [but note I'm not a lawyer so what I'm really saying is >go ask your lawyer which bring me back to my original statement that using >g++ on commercial work entangles you in copyright issues so why not just >use AT&T's [or other commercial] compiler.] I think this is total nonsense. I'm not a lawyer either, but I'd like to clarify a few things. I work for the Free Software Foundation. >I believe _main and builtin_new, and .h files at the very least, also >complicate the issue. The compiler's invisible runtime support routines are explicitly in the public domain. The only trivial .h files like <limits.h>, <float.h>, and <stdarg.h> are also in the public domain. >Even the templates by which C++ code get converted >to a particular CPU's assembly language could cloud these issues. Not so; could you copyright the word "the"? Individual instructions are below the level of what is a copyrightable entity. These objections are total nonsense. The real objection lawyers have against us is that they worry that, since we have no profit motive to keep us from doing so, we might engage in nuisance suits against people we doing things we don't like with our programs. What the lawyers fail to see, are perhaps incapable of seeing, is that there's more to life than profit. Our motive is that we *want people to use our programs*. So we're not going to go around engaging in frivolous suits that we'd lose anyway, just to bug people. Because people would stop using our programs. (Just like if, say, AT&T went around gratuitously suing people, people would stop *buying* their stuff, and AT&T would stop making money. Silly lawyers seem to think profit is a greater incentive than personal satisfaction.) >Stallman's statements not withstanding, I do not believe issue of copyright >infringement where a compiler "automatically" injects a greater or lessor >amount of its own source code into a user's work to be compiled/assembled >has been legally tested. "Fair Usage" rules certainly apply here, but what is >"fair use" when an author publicly states he doesn't really want his >work used for commercial purposes? ...I suspect the courts might find a very >narrow view of "fair use" in those circumstances. About the largest verbatim copy from the compiler sources is probably procedure prologues and epilogues, on some machines. Taken by themselves, they are obviously on the same level as the aforementioned word "the" and thus uncopyrightable. I suspect it will be a long time, if ever, before such a case makes the courts, as most people have better things to do than engage in suits they're 99.9% likely to lose. >..........In any case, if you want to use g++ for commercial work only a >competent patent/intellectual property lawyer can decide for you and your >company. Just make darned sure you and the lawyer know the intimate details >of how g++, .h files, porting to multiple machines, etc, work. 1. You can use g++ for commercial work. 2. You cannot use libg++ in commercial products without bringing them within the scope of the GNU license. We may change this in the future, if we think that is the only way GNU will be accepted into widespread use. But we will look long and hard at such a change first. Of course, you can always write a library of your own, or obtain from some other source a library that you can legally use, and compile it with g++, and use that. 3. If your lawyers worry about nuisance suits point out that we have neither the inclination, time, nor money to engage in such frivolous pursuits, and besides we're nice guys. If your lawyers still bug you, maybe it's time to get new lawyers. -- Mike Haertel <mike@stolaf.edu> ``There's nothing remarkable about it. All one has to do is hit the right keys at the right time and the instrument plays itself.'' -- J. S. Bach
ok@cs.mu.oz.au (Richard O'Keefe) (09/28/89)
In article <6602@thor.acc.stolaf.edu>, mike@thor.acc.stolaf.edu (Mike Haertel) writes: > I think this is total nonsense. I'm not a lawyer either, but I'd > like to clarify a few things. I work for the Free Software Foundation. > The real objection lawyers have against us is that they worry that, > since we have no profit motive to keep us from doing so, we might > engage in nuisance suits against people we doing things we don't like > with our programs. What the lawyers fail to see, are perhaps incapable of > seeing, is that there's more to life than profit. Our motive is that > we *want people to use our programs*. So we're not going to go around > engaging in frivolous suits that we'd lose anyway, just to bug people. > Because people would stop using our programs. Why bash lawyers? They're not here, just us apprentice wizards. Lawyers are DELIGHTED with the FSF: anyone who goes around making obscure threats (and face it, friends, the Copyleft is an obscure threat) is more likely to increase their custom than otherwise. (Just think about the number of lawyers who must have been asked "is it safe for us to use this stuff?". They get _paid_ to answer such questions.) I recently got a copy of Oaklisp from CMU. It's an object oriented Scheme variant, with multiple inheritance and all sorts of good stuff. There are lots of things I hoped to do with it, but now I'm not going to. Why? Because it's covered by an FSF-style copyleft. If I make an improvement to Oaklisp, I just simply can't afford to give it away. I can't even afford to give the diffs back to the original authors, because the CopyLeft would land me with obligations for years. I am _not_ making money out of software, it's precisely because I'm _not_ making money that I can't afford all this hassle. Another example: I picked up a copy of GPERF. It could stand a lot of improvement. But I can't afford the hassle involved in letting anyone else have the improvements, so I've thrown it away. Another example: I made a number of changes to BISON to make it work on machines without alloca(), in fact to make it work happily under System V without even needing an emulation of alloca(). (In point of fact it turns out that there was never any real benefit in using alloca() in the first place.) Have I given these diffs to anyone? No way; I'm not going to get lumbered for the next three years with the obligation to send the FULL sources to ANYONE AT ALL who asks. (That's what the CopyLeft demands, friends.) If the FSF *really* wanted to "stamp out software hoarding", they would make it easy for people to share improvements to their code. Instead, they fill every manual with 9 pages of legalistic bumf which has the *opposite* effect. -- GNUs are more derived than other extant alcelaphines,| Richard A. O'Keefe such as bonteboks, and show up later in the fossil | visiting Melbourne record than less highly derived species. (Eldredge) | ok@munmurra.cs.mu.OZ.au
eric@cfi.COM (eric) (09/28/89)
In article <6590268@hplsla.HP.COM> jima@hplsla.HP.COM (Jim Adcock) writes: >has been legally tested. "Fair Usage" rules certainly apply here, but what is >"fair use" when an author publicly states he doesn't really want his >work used for commercial purposes? ...I suspect the courts might find a very Stallman has said that he is perfectly happy to have his work used for commercial purposes. The copyleft does not prevent you from using it for any commercial product or service. You "just" have to make your source available to the customer. Maybe instead of "commercial", you meant "profitable"? :-) -- Eric Read ` | ' harvard!cfisun!eric Price Waterhouse / CFI --*-- (617) 899-6500 51 Sawyer Rd. ' | ` WUTFOTYATEOTW Waltham, MA 02154
jacob@gore.com (Jacob Gore) (09/28/89)
/ comp.lang.c++ / ok@cs.mu.oz.au (Richard O'Keefe) / Sep 28, 1989 / > Have I given these diffs to anyone? No way; I'm not going to get lumbered > for the next three years with the obligation to send the FULL sources to > ANYONE AT ALL who asks. (That's what the CopyLeft demands, friends.) A pity you subjected yourself to so much trouble based on this false premise... You DO NOT have to distribute anything to "anyone at all who asks", only to those people to whom you distribute the corresponding binaries. Jacob -- Jacob Gore Jacob@Gore.Com boulder!gore!jacob
mike@thor.acc.stolaf.edu (Mike Haertel) (09/29/89)
In article <2245@munnari.oz.au> ok@cs.mu.oz.au (Richard O'Keefe) writes: >I recently got a copy of Oaklisp from CMU. It's an object oriented Scheme >variant, with multiple inheritance and all sorts of good stuff. There are >lots of things I hoped to do with it, but now I'm not going to. Why? >Because it's covered by an FSF-style copyleft. Not "because it's copylefted" but "because it's copylefted and you obviously don't understand the terms of the copyleft." >If I make an improvement >to Oaklisp, I just simply can't afford to give it away. Sure you can, if it really *is* under the unmodified GNU license. >I can't even >afford to give the diffs back to the original authors, because the CopyLeft >would land me with obligations for years. This is complete and utter nonsense. Have you ever actually *read* the copyleft? It only takes a few minutes . . . >I am _not_ making money out of >software, it's precisely because I'm _not_ making money that I can't afford >all this hassle. Another example: I picked up a copy of GPERF. It could >stand a lot of improvement. But I can't afford the hassle involved in >letting anyone else have the improvements, so I've thrown it away. Another >example: I made a number of changes to BISON to make it work on machines >without alloca(), in fact to make it work happily under System V without >even needing an emulation of alloca(). (In point of fact it turns out that >there was never any real benefit in using alloca() in the first place.) >Have I given these diffs to anyone? No way; I'm not going to get lumbered >for the next three years with the obligation to send the FULL sources to >ANYONE AT ALL who asks. (That's what the CopyLeft demands, friends.) Ah, now we come to the crux of the issue. I can only conclude from this statement that Mr. O'Keefe has *never* *read* the license, and therefore deserves everyone's flames, for posting misinformation to the net. Excerpted for everyone's reading pleasure is the relevant portion of the license: # 3. You may copy and distribute the Program (or a portion or derivative of # it, under Paragraph 2) in object code or executable form under the terms of # Paragraphs 1 and 2 above provided that you also do one of the following: # # a) accompany it with the complete corresponding machine-readable # source code, which must be distributed under the terms of # Paragraphs 1 and 2 above; or, # # b) accompany it with a written offer, valid for at least three # years, to give any third party free (except for a nominal charge # for the cost of distribution) a complete machine-readable copy of the # corresponding source code, to be distributed under the terms of # Paragraphs 1 and 2 above; or, # # c) accompany it with the information you received as to where the # corresponding source code may be obtained. (This alternative is # allowed only for noncommercial distribution and only if you # received the program in object code or executable form alone.) # # Source code for a work means the preferred form of the work for making # modifications to it. For an executable file, complete source code means # all the source code for all modules it contains; but, as a special # exception, it need not include source code for modules which are standard # libraries that accompany the operating system on which the executable # file runs, or for standard header files or definitions files that # accompany that operating system. Notice that clause (b) applies *only* if you have made a binary distribution without source. #ifdef FLAME Mr. O'Keefe, it is my personal belief that people such as you, who inflict blatant misinformation on other readers of the net, are among the scum of the earth. #endif -- Mike Haertel <mike@stolaf.edu> ``There's nothing remarkable about it. All one has to do is hit the right keys at the right time and the instrument plays itself.'' -- J. S. Bach
ok@cs.mu.oz.au (Richard O'Keefe) (09/29/89)
ok@cs.mu.oz.au (Richard O'Keefe) wrote
: Have I given these diffs to anyone? No way; I'm not going to get lumbered
: for the next three years with the obligation to send the FULL sources to
: ANYONE AT ALL who asks. (That's what the CopyLeft demands, friends.)
In article <110004@gore.com>, jacob@gore.com (Jacob Gore) replied:
: A pity you subjected yourself to so much trouble based on this false premise...
: You DO NOT have to distribute anything to "anyone at all who asks", only to
: those people to whom you distribute the corresponding binaries.
Go read the copyleft. That's not what it says. The copyleft says
"You may copy and distribute the Program (or a portion or derivative
of it ...) provided that you ...
accompany it it with a written offer, valid for at least three
years, to give ***ANY*** ***THIRD PARTY*** free ... a complete
machine-readable copy of the corresponding source code ..."
It doesn't say "to give the people you distributed to in the first place
a complete copy" but to give ANY THIRD PARTY. That may not be what the
FSF *intend*, but it is what they *demand*. (You are allowed to pass on
the information as to where the original source may be obtained instead
of providing such a written offer, but ONLY if you didn't get the sources
yourself.) Note further that Jacob Gore MUST be wrong about this, because
if I give someone just the binaries, they are allowed to discharge their
"source" obligations by telling everyone they pass the binaries on to,
to get sources from _me_. It is clear that my "written offer" is supposed
to apply to those people as well as the people I distribute to directly.
No, if the Free Software Foundation were honestly interested in
encouraging software sharing, they would use a much simpler contract
with fewer restrictions. There is that wonderful clause
"By COPYING, distributing, or modifying the Program (or any work
based on the Program) you indicate your acceptance of this licence
to do so, and ALL its terms and conditions."
Now, picking up a copy by FTP to see what's there constitutes copying,
so this clause commits you to following every little detail BEFORE YOU HAVE
EVEN READ IT! Nice one, FSF!
--
GNUs are more derived than other extant alcelaphines,| Richard A. O'Keefe
such as bonteboks, and show up later in the fossil | visiting Melbourne
record than less highly derived species. (Eldredge) | ok@munmurra.cs.mu.OZ.aumike@cimcor.mn.org (Michael Grenier) (09/29/89)
From article <110004@gore.com>, by jacob@gore.com (Jacob Gore):
= / comp.lang.c++ / ok@cs.mu.oz.au (Richard O'Keefe) / Sep 28, 1989 /
=> Have I given these diffs to anyone? No way; I'm not going to get lumbered
=> for the next three years with the obligation to send the FULL sources to
=> ANYONE AT ALL who asks. (That's what the CopyLeft demands, friends.)
=
= A pity you subjected yourself to so much trouble based on this false premise...
=
= You DO NOT have to distribute anything to "anyone at all who asks", only to
= those people to whom you distribute the corresponding binaries.
Furthermore, you can charge a reasonable fee for the act of distribution.
You may also charge a fee to provide a warranty for the resulting new
code that you fixed.
In the worst case, you should be able to pay someone to sit a computer and
generate copies for distribution as well as running to the post office with
the disks by charging the fee as described in section 3 and under section
4.2 of the GNU Public License, terms and conditions.
-Mike Grenier
mike@cimcor.mn.orgok@cs.mu.oz.au (Richard O'Keefe) (09/29/89)
In article <6669@thor.acc.stolaf.edu>, mike@thor.acc.stolaf.edu (Mike Haertel) writes: > Not "because it's copylefted" but "because it's copylefted and > you obviously don't understand the terms of the copyleft." If that is true, and it may be, that in itself is DAMNING. > This is complete and utter nonsense. Have you ever actually *read* > the copyleft? It only takes a few minutes . . . By actual measurement, it takes me four minutes to cast my eyes over the copyleft. I have been reading and re-reading it for several days. > Ah, now we come to the crux of the issue. I can only conclude from > this statement that Mr. O'Keefe has *never* *read* the license, > and therefore deserves everyone's flames, for posting misinformation > to the net. People who can "only conclude" things which are utterly contrary to fact are at least as deserving of flames as I. I have taken care to read every word of the copyleft. I have taken care to parse every sentence. I have received six E-mail messages about this already, and before replying to each of them I have taken care to re-read every single word of the copyleft. It may well be true that I have misunderstood. I repeat, that in itself is damning. > # 3. You may copy and distribute the Program (or a portion or derivative of > # it, under Paragraph 2) in object code or executable form under the terms of > # Paragraphs 1 and 2 above provided that you also do one of the following: Here's a curious thing. The thing I'm looking at is called "version 1" of the GNU General Public Licence; it's dated February 1989; and the text Mike Haertel quotes most resembles section 4 of that document. Evidently he is not reading exactly the same document I'm reading. (The Oaklisp copyleft is still another version.) > #ifdef FLAME > Mr. O'Keefe, it is my personal belief that people such as you, > who inflict blatant misinformation on other readers of the net, > are among the scum of the earth. > #endif Mr Haertel, the information I provided was this: someone who has made an honest attempt to read the copyleft with the utmost care believes X. As a result, that person has concluded that distributing enhancements to GNU sources would be personally costly to him. If he is right, that's bad. If he is wrong, it's bad that the copyleft is so hard to read. The point at issue was not really what the FSF copyleft is intended to do, but whether it has the effect of increasing software sharing or not. In my case it has had the opposite effect. Mr Haertel can not deny this. That is a significant fact, not "blatant misinformation". It should be noted that Mr Haertel made a wild and unsupported guess about whether I had read the copyleft or not, and quoted a document other than either of the versions I was referring to. And he calls _me_ "scum of the earth". Enough already. I apologise for taking up the time of readers of this group. But I entreat you, if _you_ want to encourage software sharing, make your copyright or copyleft notices as simple and plain as you possibly can. -- GNUs are more derived than other extant alcelaphines,| Richard A. O'Keefe such as bonteboks, and show up later in the fossil | visiting Melbourne record than less highly derived species. (Eldredge) | ok@munmurra.cs.mu.OZ.au
jacob@gore.com (Jacob Gore) (09/30/89)
/ comp.lang.c++ / ok@cs.mu.oz.au (Richard O'Keefe) / Sep 28, 1989 / > ok@cs.mu.oz.au (Richard O'Keefe) wrote > : Have I given these diffs to anyone? No way; I'm not going to get lumbered > : for the next three years with the obligation to send the FULL sources to > : ANYONE AT ALL who asks. (That's what the CopyLeft demands, friends.) > In article <110004@gore.com>, jacob@gore.com (Jacob Gore) replied: > : You DO NOT have to distribute anything to "anyone at all who asks", only to > : those people to whom you distribute the corresponding binaries. > Go read the copyleft. That's not what it says. The copyleft says > "You may copy and distribute the Program (or a portion or derivative > of it ...) provided that you ... > accompany it it with a written offer, valid for at least three > years, to give ***ANY*** ***THIRD PARTY*** free ... a complete > machine-readable copy of the corresponding source code ..." >[...] > Note further that Jacob Gore MUST be wrong about this, because > if I give someone just the binaries, they are allowed to discharge their > "source" obligations by telling everyone they pass the binaries on to, > to get sources from _me_. It is clear that my "written offer" is supposed > to apply to those people as well as the people I distribute to directly. You are right, but it doesn't imply your original concern (that you personally will be tied up distributing sources for three years). Here's the paragraph that follows: ...or c) accompany it with the information you received as to where the corresponding source code may be obtained. (This alternative is allowed only for noncommercial distribution and only if you received the program in object code or executable form alone.) Only those who were given a binary-only distribution can "discharge their 'source' obligations" this way (and people doing commercial distribution are obligated to obtain and provide the source in either case). So all you have to do is only make your distribution in source form. And if you make it to a few well-known redistribution points, such as well-known ftp sites, it will get wide distribution with little cost to you (sigh, those inter-continental phone calls...) I agree, though, that reading option (b) (which you quoted) alone gives one the wrong idea. > No, if the Free Software Foundation were honestly interested in > encouraging software sharing, they would use a much simpler contract > with fewer restrictions. Offer one. > There is that wonderful clause > "By COPYING, distributing, or modifying the Program (or any work > based on the Program) you indicate your acceptance of this licence > to do so, and ALL its terms and conditions." > Now, picking up a copy by FTP to see what's there constitutes copying, > so this clause commits you to following every little detail BEFORE YOU HAVE > EVEN READ IT! Nice one, FSF! Yeah, that should probably be "By using or distributing any part of the Program you indicate...". Jacob -- Jacob Gore Jacob@Gore.Com boulder!gore!jacob
jima@hplsla.HP.COM (Jim Adcock) (09/30/89)
> Stallman has said that he is perfectly happy to have his work used for > commercial purposes. The copyleft does not prevent you from using it > for any commercial product or service. You "just" have to make your > source available to the customer. People don't seem to realize that these are two seperate businesses. Selling software, and selling source. When you sell source, you are really in the service business of hand-holding, writing documentation, running courses on how to port the software etc. It's not [necessarily] that companies don't want to sell source because it would be giving away the company jewels. Just as likely companies don't want to sell source because they cannot be profitable selling their source. Many years ago, in my business [building FFT based analysers] we decided to sell source for one of our instruments. It cost us something like $10,000 to get the order process set up and in the corporate computers. And we "sold" one source copy. --At this rate the "cost" of AT&T compilers look very cheap compared to the "true cost" of gnu compilers. The gnu position seems to be: a) Feel free to use our compilers if you make your source available for "free." and/or b) Feel free to use our compilers if you don't use our non-trivial libraries. Position a) is clearly stated [I believe] in the gnu licensing terms. Position b) is not clearly stated in the gnu licensing terms. If b) is truly an "official" gnu position, it would be nice to have this explicetely stated as part of the gnu licensing terms. Not having these things stated clearly should make your lawyer nervous. Good lawyers want things spelled out in black and white -- not shades of grey. Use whatever compiler you feel best about. Use lawyers or not as you see fit. It does not affect me any. I just don't want to see new users of C++ niavely sucked in by the "Free Software" nomiker.
mike@thor.acc.stolaf.edu (Mike Haertel) (09/30/89)
In article <2247@munnari.oz.au> ok@cs.mu.oz.au (Richard O'Keefe) writes: >Go read the copyleft. That's not what it says. The copyleft says > "You may copy and distribute the Program (or a portion or derivative > of it ...) provided that you ... > accompany it it with a written offer, valid for at least three > years, to give ***ANY*** ***THIRD PARTY*** free ... a complete > machine-readable copy of the corresponding source code ..." Once again Mr. O'Keefe opens his mouth (figuratively speaking, I guess) and nonsense comes out. This purported quote of the Copyleft omits quite a bit, including some if-then clauses that make it mean something quite different from what you see here. It appears Mr. O'Keefe is deliberately attempting to deceive the readers of this group. A previous response of mine quotes in full the relevant portion of the license; read it yourself and draw your own conclusions. I can only wonder what his motives are. I will not respond to any further postings of his. -- Mike Haertel <mike@stolaf.edu> ``There's nothing remarkable about it. All one has to do is hit the right keys at the right time and the instrument plays itself.'' -- J. S. Bach
tuck@jason.cs.unc.edu (Russ Tuck) (09/30/89)
In article <6602@thor.acc.stolaf.edu> mike@thor.stolaf.edu () writes: > >1. You can use g++ for commercial work. > >2. You cannot use libg++ in commercial products without bringing > them within the scope of the GNU license. We may change this in > the future, if we think that is the only way GNU will be accepted > into widespread use. But we will look long and hard at such a > change first. Until g++ 1.36 changed the way it links programs, this made sense to me. In effect, it says there's no copyleft on your executable unless you explicitly link in something copylefted (like libg++.a). However, g++ 1.36 now requires "-lg++" to link ANY program. So the above rules now make it impossible to use g++ without making the executable copylefted. (g++ compiles "main()" to include a call to "___main()", which is contained in libg++.a.) I consider this a bug in g++, or a dramatic policy change. (This isn't an immediate issue for me, since I'm currently at a Univ. and don't charge money when I share my code anyway. But I'm still surprised at the change.) Russ Tuck tuck@cs.unc.edu UNC Dept. of Computer Science ...!mcnc!unc!tuck CB# 3175 Sitterson Hall Chapel Hill, NC 27599-3175, USA (919) 962-1755 or 962-1932
schmidt@crimee.ics.uci.edu (Doug Schmidt) (09/30/89)
In article <2247@munnari.oz.au> you write: >Go read the copyleft. That's not what it says. The copyleft says > "You may copy and distribute the Program (or a portion or derivative > of it ...) provided that you ... > accompany it it with a written offer, valid for at least three > years, to give ***ANY*** ***THIRD PARTY*** free ... a complete > machine-readable copy of the corresponding source code ..." I'm glad you finally cited the `offending line'. Now we can be concrete. You are misquoting/misinterpreting the GPL. Here's the actual quote: ---------------------------------------- 3. You may copy and distribute the Program (or a portion or derivative of it, under Paragraph 2) in object code or executable form under the terms of Paragraphs 1 and 2 above provided that you also do one of the following: a) accompany it with the complete corresponding machine-readable source code, which must be distributed under the terms of Paragraphs 1 and 2 above; or, b) accompany it with a written offer, valid for at least three years, to give any third party free (except for a nominal charge for the cost of distribution) a complete machine-readable copy of the corresponding source code, to be distributed under the terms of Paragraphs 1 and 2 above; or, c) accompany it with the information you received as to where the corresponding source code may be obtained. (This alternative is allowed only for noncommercial distribution and only if you received the program in object code or executable form alone.) ---------------------------------------- Note the key line omitted in your posting: ---------------------------------------- You may copy and distribute the Program (or a portion or derivative of it, under Paragraph 2) in object code or executable form ^^^^^^^^^^^^^^^^^^^^^^^^^ ---------------------------------------- This section deals with the responsibilities for people who want to distribute a GNU derived *object* code or *executable* program. It states that if you want to distribute the object or executable then you also must ensure the source is available. This should be no surprise, and should not conflict with encouraging software sharing. However, this section in the GPL has nothing to do with your original complaint, which was: ---------------------------------------- I can't even afford to give the diffs back to the original authors, because the CopyLeft would land me with obligations for years. ---------------------------------------- If you can provide additional `proof' for your position I'd like to see it. It is certainly not the intention of the GPL to unduly burden those who want to help make the GNU software more robust. Doug -- schmidt@ics.uci.edu (ARPA) | Per me si va nella citta' dolente. office: (714) 856-4043 | Per me si va nell'eterno dolore. | Per me si va tra la perduta gente. | Lasciate ogni speranza o voi ch'entrate.
schmidt@zola.ics.uci.edu (Doug Schmidt) (09/30/89)
In article <9741@thorin.cs.unc.edu>, tuck@jason (Russ Tuck) writes: >Until g++ 1.36 changed the way it links programs, this made sense to me. >In effect, it says there's no copyleft on your executable unless you >explicitly link in something copylefted (like libg++.a). > >However, g++ 1.36 now requires "-lg++" to link ANY program. So the above >rules now make it impossible to use g++ without making the executable >copylefted. (g++ compiles "main()" to include a call to "___main()", >which is contained in libg++.a.) > >I consider this a bug in g++, or a dramatic policy change. It is a policy change, but not in the direction you imply above. You are referring, of course, to the gnulib3.c file with the beta release of g++ 1.36. Please look at the file right now (I assume you've got access to it, or you wouldn't be writing this ;-)). Note how this file differs from most of the other libg++ files. In fact, it looks much like the gnulib.c and gnulib2.c from the gcc release, in that it doesn't contain a copyright notice at the top. This omission means that linking with the small `compiler-helper' functions declared in this file does *not* by itself bring the executable under the copyleft (the same may not hold for other libg++ files...). In fact, certain changes to g++ 1.36 (e.g., omitting the need for a special crt0.o file) specifically mean that you *can* use the g++ compiler (and its small helper functions in the gnulib[23]?.c files) without necessarily bringing the entire program under the copyleft (assuming, of course, that you don't include the copylefted libraries from libg++, etc.). The situation described in the previous paragraph was not true with previous releases g++. That is why I said above that there was a policy change, but not in the direction you implied! I know at this point that Jim Adcock will step in and say that this whole g++/libg++ library situation is unduly confusing. I agree with him. I hope that my explanation has helped shed a little light on an often easily mis-understood issue. Doug -- schmidt@ics.uci.edu (ARPA) | Per me si va nella citta' dolente. office: (714) 856-4043 | Per me si va nell'eterno dolore. | Per me si va tra la perduta gente. | Lasciate ogni speranza o voi ch'entrate.
benson@odi.com (Benson I. Margulies) (10/02/89)
This conversation should serve as a warning to potential FSF software users. Its quite typical. The FSF hangs a legal bowl of sphagetti off of their source. Individuals, some of them trained lawyers, read this item and find that it is at best ambiguous, and at worst seems to impose all sorts of impossible conditions. Querying the FSF on this point gets a progression of results: 1) we see no problem here. The Man Himself (stallman) has said many times that his intention is to permit commercial use. 2) Oh, a statement in email of our intentions isn't good enough for you? You want some accompanying document to the copyleft, or an improvement? Well, even if there were a problem here, we aren't interested in solving it, because it would just be facilitating "software hoarding." 3) personal flames and attacks on the luckless soul who dared to raise the question. Note that the "intentions" of the FSF aren't the only relevant intentions. The copyleft, as the originator of this chain, potentially grants all kinds of rights to all sorts of others. One of them might decide to be a nuisance. -- Benson I. Margulies
davidm@uunet.UU.NET (David S. Masterson) (10/03/89)
In article <6669@thor.acc.stolaf.edu> mike@thor.acc.stolaf.edu (Mike Haertel) writes: > > #ifdef FLAME > Mr. O'Keefe, it is my personal belief that people such as you, > who inflict blatant misinformation on other readers of the net, > are among the scum of the earth. > #endif > Now hold on a second! There is no need to post this type of a flame, yet! It is very obvious that the GNU "copyleft" has a lot of interpretation and is still unclear in its intent. Both you and Mr O'Keefe quoted the same section in the GNU license and drew two different conclusions. There is a lot of good stuff coming out of FSF that people would like to make use of, but, until they feel comfortable with the "copyleft", they won't. There did not appear to be any "blatant misinformation" in what Mr. O'Keefe said, just one interpretation of a topic that has no clear interpretation yet. Flaming is the wrong approach currently -- proof of your interpretation is still needed. Note: the above does not make any statement about what the "copyleft" intends. David Masterson uunet!cimshop!davidm
jima@hplsla.HP.COM (Jim Adcock) (10/03/89)
***sigh*** ...Like I said, if you want to use Gnu software for commercial work, have what you're doing carefully checked out by a well-informed competent patent/intellectual-property lawyer. Lawyers tend to read contacts in a somewhat different light than us software types. For that matter, no matter whose compiler you're going to use, you'd be better off to run it by your lawyer before spending much time developing software on it. Better to spend a little time and money up front than run into unexpected ownership/rollayties/headaches later. It seems a sad fact in America that anyone can sue anybody over anything. Unortunately this trend is catching up with the software business. It just makes sense to find out the "truth" [according to your own lawyer] about a compiler before investing too heavily in it.
jima@hplsla.HP.COM (Jim Adcock) (10/03/89)
> Note that the "intentions" of the FSF aren't the only relevant > intentions. The copyleft, as the originator of this chain, potentially > grants all kinds of rights to all sorts of others. One of them > might decide to be a nuisance. Ignoring the rest of this response, I think this respondee touches on an important point. The legal problems one runs into in the future may not between you and your compiler supplier. Your problems might be between your company "Wayfarers, Inc." and competitor "Prune Computer Co." where Prune Computer Co. argues that they didn't infringe your product because your product was "Copyleftified" or "Public Domained" or whatevered because you stepped in the compiler agreement, and thus don't own the product you built [or something to this effect, suitably legalesed] So it doesn't help even if CompilerSuppliers, Inc. are "nice guys" who are too nice to sue you in any case. "Prune Computer Co." can still be a pain in the hindee unless you can demonstrate a perfectly clean, unbesmirched ownership of the software you write. Your competitor "Prune Computer Co." [and unscrupulous software end user company "MegaCopies, Inc"-- who might decide to copy your software without paying you royalties] are the real people to worry about. You've just got to be able to demonstrate unbesmirched ownership to your software product. Muddy license agreements to compilers and/or libraries defeat this.
bruce@blender.UUCP (Bruce Thompson) (10/04/89)
You know, it's funny, but everyone seems to be bitching about apparent obligations that the FSF *REQUIRE* you to be under simply for obtaining Software from them. Let's look at the first sentence where people having been quoting: "You may distribute....." ^^^ This does not say: "You MUST ...", it leaves it open! IF YOU CHOOSE to distribute the software, THEN and ONLY THEN are you placed under obligations. IF YOU CHOOSE NOT TO DISTRIBUTE the software THEN YOU ARE UNDER NO OBLIGATIONS OF ANY KIND! So, if you make modifications which you do not distribute, then you don't have to do anything! At the same time, if you make modifications, you can at your option, distribute the modified version, with full source, etc. OR you can distribute the original you received, with full source, etc. The CONCEPT we're dealing with here is simple. The LEGALEZE which expresses this concept is COMPLEX. This should not be all that supprising, as legal documents are intended to precisely define the term of a contract, which may require a lot of apparently useless verbiage. Hope this helps, Bruce. ------------------------------------------------------------------------ Bruce Thompson CAE Engineer, NovAtel Communications Ltd. 1020 64 Ave, N.E., Calgary, Alberta, CANADA Disclaimer: I do not represent in any way the views/policies of my employer, and often not even my own. "There is no conclusive evidence of life after death. But there is no evidence of any sort against it. Soon enough you will KNOW. So why fret about it?" - Lazarus Long
eric@cfi.COM (eric) (10/04/89)
In article <6590269@hplsla.HP.COM> jima@hplsla.HP.COM (Jim Adcock) writes: >eric@cfi.COM (Eric Read) writes (I think, attribution missing) >> Stallman has said that he is perfectly happy to have his work used for >> commercial purposes. The copyleft does not prevent you from using it >> for any commercial product or service. You "just" have to make your >> source available to the customer. > >People don't seem to realize that these are two seperate businesses. >Selling software, and selling source. When you sell source, you are really >in the service business of hand-holding, writing documentation, running >courses on how to port the software etc. It's not [necessarily] that Whoa! Where did you find this in the GNU license?? Just do what the FSF does - ship them a valid source tape, cash their check, and show them the door. If they want services, tell them there are plenty of consultants around. As long as it compiles to the same binary you sold them, your obligation is met. -- Eric Read ` | ' harvard!cfisun!eric Price Waterhouse / CFI --*-- (617) 899-6500 51 Sawyer Rd. ' | ` WUTFOTYATEOTW Waltham, MA 02154
jima@hplsla.HP.COM (Jim Adcock) (10/06/89)
>>People don't seem to realize that these are two seperate businesses. >>Selling software, and selling source. When you sell source, you are really >>in the service business of hand-holding, writing documentation, running >>courses on how to port the software etc. It's not [necessarily] that > >Whoa! Where did you find this in the GNU license?? > >Just do what the FSF does - ship them a valid source tape, cash >their check, and show them the door. > >If they want services, tell them there are plenty of consultants around. >As long as it compiles to the same binary you sold them, your obligation is met > >Eric Read Some companies believe in supporting what they sell. Many companies view their "obligations" as something different than [and conflicting with] the FSF "obligations requirements." Presumably, these companies would rather pay for a compiler than change their corporate philosophies of what "obligations" they feel they owe their customers. Can you give any examples where companies have made major changes in their business practices in order to be able to use FSF compilers? NeXT maybe? Are FSF licensing restrictions actually changing how software companies do business, or are they just tilting at windmills? Sometimes a company's view of their obligations make for some strange scenerios. Thus HP assists FSF in porting their products to HP machine, in the hope this will benefit some HP customers, while at the same time being frustrated by FSF licensing requirements from using the compilers ourselves.
ok@cs.mu.oz.au (Richard O'Keefe) (10/06/89)
In article <110@blender.UUCP>, bruce@blender.UUCP (Bruce Thompson) writes: > You know, it's funny, but everyone seems to be bitching about apparent > obligations that the FSF *REQUIRE* you to be under simply for obtaining > Software from them. You have completely misundertood the debate. The point of the debate is that fear and uncertainty (however unwarranted) about the copyleft has prevented people from contributing to the improvement and dissemination of FSF programs, which defeats the aims of the FSF. > IF YOU CHOOSE NOT TO DISTRIBUTE the software THEN YOU ARE > UNDER NO OBLIGATIONS OF ANY KIND! This is not true. The copyleft makes it clear that merely to make a copy of the software constitutes accepting all the conditions of the copyleft: "By COPYING, distributing, or MODIFYING the Program (or any work based on the Program) you indicate your acceptance of this licence to do so, and ALL its terms and conditions." Because I changed Bison so that it didn't need alloca(), that obliged me to obey all the terms and conditions of the copyleft. The whole point of the debate was that my fear and uncertainty about the other terms meant that I did not dare to do anything with my changes except sit on them. But the FSF is opposed to "Software hoarding"! There is considerable legal uncertainty about what constitutes distribution. If I have FSF object code on my system, and someone makes a copy without my permission, have I "distributed" the code? How hard must I try to prevent unauthorised access? If I run a mail relay site, and someone obtains an FSF program by mail, with the mail being relayed through my site, have I copied or distributed the program? If I keep a copy of FSF sources in a public FTP area on my machine, and someone makes changes to those sources without my knowledge (and in doing so fails to "cause the modified files to carry prominate notices stating" who changed the files and when), and a third party picks up a copy of the modified files, am _I_ in violation of the copyleft? If I were running a machine which permitted general access via FTP or uucp, I would not allow any FSF sources on that machine until I had _written_ confirmation from the FSF that they knew what precautions I intended to take against unauthorised modifications and were satisfied with them. I don't for one moment suggest that they would demand anything unreasonable, but with the present form of the copyleft there is no point in taking risks.
eric@cfi.COM (eric) (10/10/89)
In article <1989Oct2.151317.3789@odi.com> benson@odi.com (Benson I. Margulies) writes: >This conversation should serve as a warning to potential FSF software >users. Its quite typical. The FSF hangs a legal bowl of sphagetti off Ambiguity. Fear, uncertainty, doubt. Frustration. It is true that it doesn't help to argue about this - the existence of so much argument demonstrates the problem. Can any of us think of a more "comforting" copyleft? As a new legal concept, it is bound to take more than a couple of iterations (not to mention some actual case law) before any normal lawyer will let it pass. What can we suggest to improve the situation? Are there any analogous legal scenarios that can be invoked (precedent is everything in the law)? Or are for-profit activities forever constrained from working with FSF software by the need to control risks and exposure? Certainly the FSF is losing a lot of their potential effectiveness as a result of our uncertainty - do they care? Well, a lot of questions from someone going on vacation tomorrow. :-) -- Eric Read ` | ' harvard!cfisun!eric Price Waterhouse / CFI --*-- (617) 899-6500 51 Sawyer Rd. ' | ` WUTFOTYATEOTW Waltham, MA 02154