rms@AI.MIT.EDU (06/06/89)
Please don't publish statements about the GPL or what it implies
unless you understand it properly and quote it correctly. I suggest
that you (1) ignore what other people have said on this subject on
info-gcc, since most of them are wrong, and (2) verify your
understanding privately with an expert before you express an opinion
based on that understanding.
Here a person (who is not, in general, flaming) intends to restate
our policies to criticize them:
Any software developed using Gnu software, ...
..., which is distributed to others must be
available to them under the terms of the GPL: Free, with
source, and without any restrictions on redistribution.
...BUT - it also seems to mean that if I use gcc to compile my software
then my software is covered by the GPL.
This is misinformation. We have never said that software "developed
using" GNU software is covered by the copyleft. There is nothing
in the GPL about this. It covers only copying all or part of a
GNU program.
(Such a condition on use would be justifiable ethically, since
hoarding doesn't deserve help. But it would be bad tactics, and
copyright law wouldn't let us enforce it. Anyway, we have already
given permission for such use in the GPL, so it is too late to change
our minds.)
He continues,
i.e., I am including Gnu
header files and libraries into my code.
This is another inaccuracy: it says in effect that using GCC is
synonymous with including GNU header files and libraries. In fact,
those are two entirely independent activities. A compiler is one
thing, and libraries are another, and header files are another.
The only exception to that is the small header files and one library
used automatically with GCC. However, these are in the public domain,
as you can see if you look at them. As I've said many times, I felt
those were too small to make sense as a lever to persuade someone to
share a program.
The other GNU libraries are copylefted, and as a result, you can't
distribute an executable linked with them unless you are also
willing to distribute the rest of that program under the copyleft.
However, ordinary use of GCC would not result in use of these
libraries, unless you install them in libc. You would have
to go out of your way to do such a thing.
I am not certain that including a header file of small definitions is
covered under the copyleft at all. Including a file containing
functions, or macros generating sizeable pieces of code, would count;
but it isn't clear whether including a typical header results in
copying anything significant into the resulting object file.
THIS IS THE REASON THAT I (as a seller of commercial software) DON'T
DARE USE GNU SOFTWARE in any way relating to the products that I ship.
It is because of this threat of "contamination". [1]
As far as I know, the term "contamination" has no referent in this
situation. Either your program contains a significant piece of GNU
code or it doesn't. Its history doesn't matter.denny@mcmi.UUCP (Denny Page) (06/07/89)
rms@AI.MIT.EDU writes: >The other GNU libraries are copylefted, and as a result, you can't >distribute an executable linked with them unless you are also >willing to distribute the rest of that program under the copyleft. 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? -- Someday has arrived
brooks@vette.llnl.gov (Eugene Brooks) (06/07/89)
I don't want any hate mail from the GNU staffers for this, I thought I would answer this question as it covers an interesting point in copyright law. I support the GNU effort with my time now and then and would certainly send a check to FSF if they needed to defend their copyleft. In article <1108@mcmi.UUCP> denny@mcmi.UUCP (Denny Page) writes: >rms@AI.MIT.EDU writes: >>The other GNU libraries are copylefted, and as a result, you can't >>distribute an executable linked with them unless you are also >>willing to distribute the rest of that program under the copyleft. > >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? This is the one certified way of beating the copyleft, which RMS does not like to advertize. It even has the feature that the copyleft prevents the end user from redistributing the resulting binary because he does not have part of the source. Its an interesting turnabout, the copyleft is protecting the resulting proprietary program. Now, you might think that FSF would never sue an end user under these conditions. No one would think that FSF would sue an end user for distributing proprietary code freely. However, an interesting feature of the copyright law is that if you don't defend your copyright you lose it. If you allow someone to freely distribute your goods in violation of the copyright (as in the case of the end user noted above), and then a company distributes the goods in violation of the copyright in "software hoarding mode", FSF will not successfully defend the copyleft in court against the second party unless they defended it against the first. This means, in effect, that FSF MUST defend your proprietary source from redistribution if it is to keep its copyleft in force! brooks@maddog.llnl.gov, brooks@maddog.uucp