raulmill@usc.edu (Raul Deluth Rockwell) (12/22/89)
In many articles, [various people] write: >>> If I link with GNU I can't redistribute it the way I want. >> So link with your own code. > Ick. If there really is such an incredible demand for unencumbered linking libraries, it seems like sooner or later someone will write some 8^) However, since there are a number of good reasons for 'later' rather than 'sooner', and since there is an incredible volume of slightly heated traffic on the subject, here is my immodest proposal: (1) Design an elegant set of linking libraries which can be used in the place of GNU's. (2) Write them. (3) Work out an acceptable set of distribution terms (e.g. freely redistributable, $100000 per copy license fee, only on alternate Tuesdays, whatever...). (4) Distribute them. The current law of the land is, roughly: "He who writes the code gets to make the distribution rules." I like it that way, and I expect most of you do as well. -------------------------------------------------------------------- Aside: That linked libraries must be distributed with code is as least as much a technical detail as it is a moral issue. -- Raul Rockwell INTERNET: raulmill@usc.edu ! UUCP: ...uunet!usc!raulmill ! 55 mph = 82 nc U.S.SNAIL: 721 E Windsor #4, GLENDALE CA 91205 ! --
raulmill@usc.edu (Raul Deluth Rockwell) (12/23/89)
In article <RAULMILL.89Dec21114548@aludra.usc.edu> raulmill@usc.edu (Raul Deluth Rockwell) writes:
[... cancel that, you didn't really want to read that stuff
anyways, did you? ...]
A short time after I made that posting, somebody wrote me to say that
I am already too late, that there is already at least one group
working on creating freely redistributable ANSI/POSIX linked libraries
for C. Oh well 8^)
Of course, these routines aren't all written yet, nor are they ready
for distribution, which makes them almost exact (drop-in 8^)
replacements for the GNU libraries. That ought to make Everybody happy!
[meanwhile, back at the ranch]
The whole issue of distribution of linked libraries seems a bit
short-term to me. It only applies to linking prior to distribution (a
rather arcane concept). There is also installation-time linking (I
seem to remember NEXT using this as a belated workaround of GPL) and
run-time linking (I don't suppose you've ever heard of the Amiga? No?
How about Multics? No? Maybe SunOS (v4)? No? I'm sure there are
other examples 8^).
I can just see it: $100000000 lawsuit charging for infringement of
copyright, seems somebody wrote some libraries for C which are
identical to somebody elses. Happens that both are a set of calls to
a dynamic (run-time) library.
You can't prevent silliness in everybody, at times you can only work
around it.
--
Raul Rockwell
INTERNET: raulmill@usc.edu !
UUCP: ...uunet!usc!raulmill ! 55 mph = 82 nc
U.S.SNAIL: 721 E Windsor #4, GLENDALE CA 91205 !
--chase@Ozona.orc.olivetti.com (David Chase) (01/11/90)
In article <RAULMILL.89Dec22084945@aludra.usc.edu> raulmill@usc.edu (Raul Deluth Rockwell) writes: >A short time after I made that posting, somebody wrote me to say that >I am already too late, that there is already at least one group >working on creating freely redistributable ANSI/POSIX linked libraries >for C. Oh well 8^) > >Of course, these routines aren't all written yet, nor are they ready >for distribution, which makes them almost exact (drop-in 8^) >replacements for the GNU libraries. Actually, I'm fairly certain that there's various implementations available under various circumstances. For instance, I know of one ANSI-C library that was portable and functioning (based on some draft from X3J11) about mid-way through 1988, but since its ownership was not clear, it was judged not worth the effort to try to do anything with it. But, they are real, and even come with their own test suite. David