Path: gmd.de!Germany.EU.net!mcsun!uunet!europa.eng.gtefsd.com!emory!
darwin.sura.net!zaphod.mps.ohio-state.edu!pacific.mps.ohio-state.edu!
cis.ohio-state.edu!gnu.ai.mit.edu!rms
From: r...@gnu.ai.mit.edu (Richard Stallman)
Newsgroups: gnu.misc.discuss
Subject: Why we decided to replace compress
Date: 16 Apr 1993 00:33:26 -0400
Organization: GNUs Not Usenet
Lines: 43
Sender: dae...@cis.ohio-state.edu
Distribution: gnu
Message-ID: <9304160433.AA07470@mole.gnu.ai.mit.edu>

The LZW compression algorithm is covered by a patent owned by Unisys,
and by another patent owned by IBM.  If you distribute compress,
either of these companies might decide to sue you tomorrow.  Or
perhaps they won't.

The FSF does not want to put users in this sort of vulnerable
situation.  Our mission is to give users programs they can freely
share.  Which means, implicitly, that they can share without risk of
being sued by IBM or Unisys.

It took two years of looking to find a non-patented compression
program that's comparable with compress.  It happens to be better,
technically, but that is a bonus.  The important thing is that we are
no longer vulnerable--at least from those particular patents.  (One
thing about patents is that you can *never* be sure you are safe.)

We could have kept our heads in the sand and told ourselves, "We
haven't been sued, so everything's ok."  Then, if one day IBM or
Unisys did threaten us (or threaten users, which would have a similar
effect), it would have been too late to do anything.  We could not
have found a replacement program on short notice.  The whole community
would have been in a jam.

Instead, we took warning from threats stated to others, such as the
POSIX standards committee, and found an alternative *before* the
problem happened.

I'm not surprised patent holders advise people to relax until they
themselves are actually attacked.  No sensible marauder wants
potential victims to take precautions in advance. :-)

We were lucky to find a way around this particular software patent.
That leaves about 6000 other software patents buried who-knows-where,
ready to explode if we step in the wrong place.  You are walking
around in the same minefield.  Perhaps the code you wrote today does
not infringe a patent--but what about tomorrow?

So don't wait till you step on a patent yourself.  Anticipate the
problem, join the League for Programming Freedom, and help turn off
the mines before any more people step on them.

Look in /doc/lpf on ftp.uu.net for files giving information, or send
mail to l...@uunet.uu.net.

Path: gmd.de!Germany.EU.net!news.dfn.de!darwin.sura.net!
zaphod.mps.ohio-state.edu!pacific.mps.ohio-state.edu!cis.ohio-state.edu!
gnu.ai.mit.edu!rms
From: r...@gnu.ai.mit.edu (Richard Stallman)
Newsgroups: gnu.misc.discuss
Subject: Why we decided to replace compress
Date: 16 Apr 1993 01:09:38 -0400
Organization: GNUs Not Usenet
Lines: 25
Sender: dae...@cis.ohio-state.edu
Distribution: gnu
Message-ID: <9304160509.AA07537@mole.gnu.ai.mit.edu>

If the LZW patents did apply only to hardware implementations, they
would not be software patents, and we would be safe using compress.

However, Unisys sent a threatening letter to the POSIX standards
committee when the committee considered specifying compress as part of
the POSIX standard.  The letter said anyone implementing the LZW
compression algorithm in a POSIX system would have to pay Unisys.  It
was sent in 1990 or 1991.

This threat clearly was mainly meant for software implementations.
(Which would most likely be compress itself.)  While someone might
implement LZW compression in a POSIX system using special-purpose
hardware, this wasn't likely to happen often.

If Unisys officially said at some point that it would exempt software
implementations and demand royalties only for special-purpose
hardware, then perhaps this could be held against them in court.  But
that still leaves the IBM patent, which is a separate threat.  There
is no reason to think it applies only to hardware implementations;
when I discussed it with IBM people (including Wegman, who is one of
the patent applicants, and a lawyer), we took for granted that it
applied to software implementations, and they never said anything to
the contrary.