Not logged in
Log in now
Create an account
Subscribe to LWN
LWN.net Weekly Edition for May 23, 2013
An "enum" for Python 3
An unexpected perf feature
LWN.net Weekly Edition for May 16, 2013
A look at the PyPy 2.0 release
Good luck doing development without RMS's compiler.
First FOSS OS?
Posted Mar 22, 2007 10:43 UTC (Thu) by NigelK (guest, #42083)
Posted Mar 22, 2007 18:26 UTC (Thu) by JoeBuck (subscriber, #2330)
Posted Mar 22, 2007 18:44 UTC (Thu) by NigelK (guest, #42083)
I didn't say it was before GCC, I said before I'd ever heard of GCC.
I don't remember the name - it was something I bought from a PD (Public Domain) library on floppy 15 to 20 years ago. It was a DOS program, supported the full C spec at the time (or as much of it as I used, apart from fork() obviously), and allowed me to write graphical programs for MSDOS5.
It was at least free as in beer, although I don't remember if the source was included or not. I stopped using it in 1992 when I started programming HPUX workstations.
The point is that if that FSF hadn't existed, if RMS didn't write GCC, someone else's C compiler would have taken its place as the compiler of choice simply because other people were writing those sorts of thing at the time anyway. The Public Domain scene was *massive* for a time during the 80's and early 90's before Internet access became more widespread.
Posted Mar 22, 2007 21:23 UTC (Thu) by landley (guest, #6789)
Anybody remember Jim Butterfield? My commodore 64 came bundled with a
bonus disk of software, about half of which was public domain stuff he'd
written. Here's an interview with him that predates the founding of the
FSF by a year, where he talks about writing public domain software and
author's rights to do what they like with their programs:
Posted Mar 23, 2007 13:01 UTC (Fri) by filker0 (guest, #31278)
Posted Mar 23, 2007 14:18 UTC (Fri) by NigelK (guest, #42083)
You can thank Unipress for this, I think. If I recall correctly, they're the ones that pushed RMS over that cliff.
Whilst it was RMS who came up with the GPL itself, it's another case of there would have been someone else to step into the breach had he not done so. There were lots of rumblings during the late-80's (GPL1 was written in 1989) over these same issues, especially credit-stripping. Some people by that time were already putting their own usage and distribution restrictions in their README and/or documentation files. RMS was probably the first one to came up with a popular set of restrictions, but my licence history isn't so good here.
But if RMS had decided to go into gardening instead, we'd still be pretty much where we are today, just with different players but similar licences and projects.
Free software, FOSS, RMS, and too much discursion
Posted Mar 23, 2007 15:38 UTC (Fri) by filker0 (guest, #31278)
About that same time, AT&T decided to try to make a commercial go at Unix, and the licensing
became much more restrictive (and expensive) than it had been. This made it more difficult for
software tinkerers to try out new OS ideas within an existing OS. (No other "full featured" OS was
available on so many different hardware architectures at the time.) This is where the "Pubix"
project came from, and a motivation for GNU (the original goal) as well. Keep in mind, this was
in the early to mid 1980s. The PC was not the target of most of these earlier projects; PDP-11s,
VAXen, other mini-computers, plus DEC-10/20s, and most likely Burroughs, Univac, IBM, CDC
and Honeywell mainframes, were the target systems. It wasn't until the 386 that GNU software
really had much penetration into the Intel based personal computer segment.
My motivation for getting into the free software community was a mix of expansion of
knowledge, sharing of ideas, support for odd environments and architectures, and to improve
the quality of software. I've never had the anti-commercial software / anti-copyright agenda,
though I am firmly and vocally anti software patents. I agree with Linus about GPLv2, and think
that GPLv3 is more a political statement than it needs to be. I think that the anti-* stuff that has
been added to the draft should be in "optional" parts that can be included or excluded as the
author of the original work chooses. My original view of free software was that if you incorporate
a free software component into your proprietary work, anyone should have the right to the
source for that component in its original form and, in most cases, your modification of it, but did
not extend to the entire composite that included that component. I saw some of the early GPL
stuff and thought it was over-reaching, and continued to release code into the public domain
well into the 90s before I ever released anything under the GPL. The library license made me
I knew RMS at the time all this was happening, as he and I shared some friends. I doubt that he
ever took notice of me; I did help Martin Minow maintain Dave Conroy's DECUS C collection (I did
the POS version of libc, and contributed in one or two other minor places), and he invited me to
parties that Stallman also attended. I was not overly impressed with RMS, but then again, I never
worked with him nor dealt with him on any technical issues, just met him at parties in the Boston
area, and he tended to be a bit arrogant, and I was somewhat arrogant myself in my own quiet
way and found him a bit abrasive. It was a long time ago.
My heros in the Free Software arena include the afore mentioned David Conroy (now at
Microsoft), Martin Minow (greatly missed), Andrew S. Tannenbaum (or however he spells it), Larry
Wall (even though I've never liked Perl much), along with a few others who nobody reading this
are likely to have heard about and that I've lost track of.
RMS may still be involved in the development of software, but he has made himself into a full-
time fundimentalist Free Software evangelist. His stated positions are too absolute for my tastes,
but I believe that if he were to bend towards any middle ground, his position as avatar of the
Free Software movement will be forever compromised. You might want to view RMS as the Pope,
with ESR as Martin Luther. Me? I'm Jewish, so I have my own beliefs. At least neither Richard nor
Eric has ever tried to imprison me as a non-believer, though I think that Eric may have once
joked about doing me harm (at a filksing; I don't remember for sure) for a bad pun or some such.
I am not anti-RMS, but I'm not pro, either. I do believe that the GNU project has done a lot of
good. I have personally benefited from it. I believe that GNU and FSF have a place in today's
market of ideas. I do not believe that they should be able to retroactively add new restrictions to
existing GPL licensed software that they don't own; what they do with the stuff they do own is
their business. I am anti-DRM, anti-software patent, and anti-copyright abuse.
So getting back to the original point -- RMS was pushed over the edge to become the advocate
he became because of what happened with one of his creations, emacs. Had it not been him, it
would have been someone else. It took many years between the announcement of the GNU
project and the rise of FSF in the free software movement. GCC became what it is now because
hardware vendors saw it as a way to get a compiler for their new processor on the cheap, so they
helped with its development, either funding others to do it or doing the work themselves and
releasing that back "upstream" (as they were required to do).
I will not belittle what RMS has done, but I don't believe that he's the pioneer that some paint him
as (I'm not sure that he paints himself that way, I've not spoken to him personally since 1985).
He is a figure that looms large, and has a personality to go with it. He is no closer to the truth
than many others, but he's louder than most. He stands on the shoulders of giants, true, but
he's not alone, and others stand on his shoulders. Don't discount him; He has a lot of influence.
Don't take him too seriously; that leads to a narrowing of your own horizons. Think for yourself
and don't be afraid to agree with him on some things and disagree on others.
Posted Feb 9, 2011 7:42 UTC (Wed) by rs79 (guest, #72801)
Vik Sondi and I were the first two people I know of to use it, presumably others at Waterlook did, although I went there the next year and met most of the Unix poeple and never saw Dave's compiler used there - we didn't need to we had real Bell Labs Unix on a PDP 11/45 and it *had* a C compiler already - they had no need of RSX there. And we sure coldn't use Unix for real work. And RT-11 Sucked. Badly.
Dave gave the compiler to DECUS, the DEC users group. It popped up on my radar in Los Angeles in 1984 when we used it to generate Z8000 code. I noticed Dave's name was still in it. By 1992 or so I'd noticed the free C compiler I got from John "hoptoad" Gilmore's site said it was a port of the DECUS C compiler - and was now called "gcc".
I dunno what RMS wrote, but from what I've seen Dave Conroy wrote what is now called gcc.
As an aside, I met Charles Forsyth briefly at Waterloo. He looked a bit like a grown up Harry Potter and always wore a blue blazer. Dave looked like a deadhead, they were quite the pair.
There's a picture of me and Dave at the 1975 Teklogix company picnic in Terra Cotta, Ontario here: http://rs79.vrx.net/works/photoblog/2011/Feb/9/.rjsdgcs.j... I'm in the bicycle getup, I'd ridden 50 miles to get there, Dave is in the blue shirt and long hair.
Posted Mar 22, 2007 17:30 UTC (Thu) by landley (guest, #6789)
Fabrice Bellard's Tiny C Compiler (http://tinyc.org) is somewhat stalled
since all his time's going into qemu these days, but last I checked the
wikipedia entry for it (http://en.wikipedia.org/wiki/Tiny_C_Compiler)
still linked to the mercurial repository of my personal fork of the
project (http://landley.net/hg/tinycc) as a way to stay up to date. I
should probably also update
http://fabrice.bellard.free.fr/tcc/tccboot.html when I get the time...
So yeah, I've got that covered. It's still just on my todo list because
I can still get GCC under GPLv2, but when that goes GPLv3 only I've got a
I've used Turbo/Borland C, I've used Watcom's x86->PPC cross compiler,
I've used IBM's VisualAge, I've used Sun's C compiler on Solaris, and
probably a few others I'm not remembering at the moment. Nobody ever
called OS/2 Watcom-OS/2, or called WWIV Turbo/WWIV. Using that as
justification for the GNU/Linux/Dammit crusade is deeply silly. Even
Microsoft doesn't claim to own programs built with Visual C or Quick C
(most of the time).
Posted Mar 22, 2007 18:04 UTC (Thu) by nix (subscriber, #2304)
Of course GCC will never go GPLv3 only: that would have catastrophic consequences, as no non-GPLv3ed software would be buildable with such a GCC (which is obviously deleterious). The runtime libraries will remain under various flavours of GPL+exception (the precise nature of the exception varying depending on the nature of the language so as to ensure that programs compiled with GCC are not encumbered: the GNAT runtime, with its cross-unit inlining, needs different exceptions from libstdc++ needs, which needs different exceptions to libgcc or libjava...)
The *compiler* will probably go GPLv3 (except for the docs *sigh*), but GCC is more than just a compiler.
Posted Mar 22, 2007 18:28 UTC (Thu) by JoeBuck (subscriber, #2330)
Posted Mar 22, 2007 21:09 UTC (Thu) by landley (guest, #6789)
They came up with lgpl for glibc because significant chunks of the
library wind up copied verbatim into the resulting program (especially
when statically linked), so you can make a strong case that it IS a
derived work. But translation software shouldn't slap an extra layer of
copyright on someone's document when it turns french into spanish.
There's no additional creative element embodied in the resulting work.
Having the exception to make you feel better is one thing, but compiling
a program with Microsoft's proprietary compilerm, or Sun's Java compiler,
doesn't make the result owned by Microsoft either. (The runtime
libraries are another matter.)
Posted Mar 23, 2007 14:19 UTC (Fri) by nix (subscriber, #2304)
GCC *does* textually copy parts of itself (libgcc, libstdc++ headers,
<stddef.h>...) into programs built with it (some of these technically get
copied by the linker, but it's GCC that induces the linker into doing that
Microsoft *do* claim rights over programs you build with their C compiler,
again because of the language runtime (mostly? they may have other
patent-related reasoning which I'm not really interested in since it
doesn't apply to anything remotely free).
Posted Mar 23, 2007 14:16 UTC (Fri) by nix (subscriber, #2304)
Copyright © 2013, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds