NetNews Usenet Archive 1992 #30/NN_1992_30.iso/spool/comp/os/linux/20609

Internet Message Format | 1992-12-16 | 3.8 KB

Xref: sparky comp.os.linux:20609 comp.unix.pc-clone.32bit:821 comp.unix.wizards:5245 comp.unix.misc:4595 gnu.misc.discuss:3993
Path: sparky!uunet!stanford.edu!snorkelwacker.mit.edu!ai-lab!life.ai.mit.edu!friedman
From: [e-mail] (Noah Friedman)
Newsgroups: comp.os.linux,comp.unix.pc-clone.32bit,comp.unix.wizards,comp.unix.misc,gnu.misc.discuss
Subject: Re: ANNOUNCEMENT: Alpha release Linux/GNU/X unix clone on CDROM for PCs
Followup-To: gnu.misc.discuss
Date: 16 Dec 92 00:53:53
Organization: Free Software Foundation, 675 Mass Ave. Cambridge, MA 02139
Lines: 52
Message-ID: <[e-mail]>
NNTP-Posting-Host: nutrimat.gnu.ai.mit.edu

Stallman said he thought it would only take 2 years to finish the entire
system.  He underestimated a lot; he obviously didn't take Hofstadter's
Rule into consideration (for all the good that would have done :-)).

In the 8 years since the start of the GNU Project, a considerable amount of
software has been written.  This software has been maintained by either a
small number of staff with many simultaneous projects or by volunteers who
did what they could when they had time.  In addition, more time has been
spent porting and merging ports of GNU software to different platforms than
had been originally intended.  I think this has been of considerable
benefit to the FSF (such as increasing the number of people who use GNU
software and support us with contributions), but it has also slowed down
development of new software.

We waited several years for Mach to become free because it seemed like the
best system to use; several alternatives (such as Trix and Sprite) were
considered, but none of them offered the functionality we wanted, and
writing a microkernel ourselves would have been foolish.  A lot of thought
and experience which we simply lack have gone into the development of Mach.
For a couple of years it seemed like the release of Mach was imminent, so
we kept waiting for it to become free (any development of the Hurd would
have been a wasted effort otherwise).  Now mach has been made free and for
the last two years the Hurd has been developing at a reasonable pace.

Because it's taken so long to get a complete system out the door, some of
our goals have already been partly achieved by other groups as well; for
example, CSRG has managed to make most of BSD free (assuming USL doesn't
win their case), Linux is available for 386 systems, and so on.  An
increased amount of free software is a good thing.  But it doesn't obviate
the need for GNU.  GNU will be complete (unlike net2), and Linux only works
on one hardware platform.  And none of these systems offer some of the
features we plan to implement.  By the way, many other operating systems
use GNU software (the C compiler is a particularly common example).
Without the GNU Project the release of their software might have taken
longer.

Proprietary systems like OSF/1, Windows/NT, and so on may be useful
operating systems, but they fail to meet a fundamental goal of the GNU
Project---they are not free---so they do not make GNU obsolete.

The GNU Manifesto expresses Richard Stallman's feelings as of 1985.  He
still personally prefers that all software be free one way or another.  The
FSF is a different story, though.  It frowns on proprietary software, but
its mission is to promote free software, not to act directly against
proprietary software.

The Manifesto suggests one method of providing funding for development of
free software, a "software tax".  This was just one possible suggestion by
Stallman.  The FSF does not lobby for any such legislation.  However,
similar taxes already exist.  The NSF, NIH, and other government bureaus
frequently give grants or pay the salaries of university employees who
conduct research and/or develop various things, including software.