🚀 go-pugleaf

RetroBBS NetNews Server

Inspired by RockSolid Light RIP Retro Guy

Article View: gnu.misc.discuss
Article #74461

Re: LZW compression patent

#74461
From: Dan Bernstein
Date: Tue, 07 May 1991 02:50
57 lines
3130 bytes
In article <1991May3.132441.20468@cs.utk.edu> Dave Sill <de5@ornl.gov> writes:
> What exactly does the Unisys patent cover?  Presumably, compression
> using LZW is protected.  How about the format of compressed files?

The Unisys patent is available from uunet in pub/lzw-patent.Z. It's
perfectly readable. By all accounts it covers compression using LZW.
However, it was preceded by the IBM Miller-Wegman patent, which also
appears to cover LZW.

> How about uncompression, though?

The way the patent is worded, only a ``data compression and data
decompression system'' is covered. Unfortunately, ``system'' is a vague
term. If you compress a file, and someone else decompresses it, a lawyer
who argues that the entire process forms a ``system'' will probably win.
The IBM patent is much more straightforward.

> The concern is that if Unisys were
> to enforce its patent against compress/uncompress/zcat, would there be
> a danger of not being able to uncompress previously compressed files?

In theory, yes, but in practice nobody will care if you uncompress a
file one last time and store it in a different format.

> What is FSF's migration strategy wrt compress?

You'll have to ask Richard Stallman. He has been offered at least my
high-compression AP and Y coders and Ross Williams' high-speed LZRW1
routines; Y and LZRW1 are, to my knowledge, free of all restrictions.
So he certainly doesn't lose anything by including them in GNU. I don't
expect, though, that he'll immediately stop using compress, as most
sites do not have the new programs.

> I notice they're still
> distributing compress and compressed files on prep, and hence are
> liable to charges of patent infringement.

Probably not. The patents basically only protect against use of the
compressor and decompressor; keeping a file around in compressed form
does not infringe. It might contribute to someone else's infringement,
though, and contributory infringement is actionable if the direct
infringement can be proven.

> Is yabba/unyabba going to
> replace compress?

There's no way to tell yet. Individual links can easily switch to a
different compression method, but archive sites will stay with compress
until a better standard is widespread. (Rather, perceived to be
widespread.)

One thing an archive site can do is make selected files available in
both .Z and .Y form, and see how often the .Y version is retrieved. If
in a few years most people take the smaller file, the site might try to
drop .Z and see what happens.

---Dan
Technical Details
Message-ID: <4642:May702:50:2691@kramden.acf.nyu.edu>
Path: arc1-ca.pugleaf.net!gmdzi!unido!mcsun!uunet!zaphod.mps.ohio-state.edu!think.com!hsdndev!cmcl2!kramden.acf.nyu.edu!brnstnd
References: <1991May3.132441.20468@cs.utk.edu>

Show Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit
X-Google-Thread: 1025b4,8cf944b40b9f1167
X-Google-Attributes: gid1025b4,public
X-Google-ArrivalTime: 1991-05-07 14:11:43 PST
Path: arc1-ca.pugleaf.net!gmdzi!unido!mcsun!uunet!zaphod.mps.ohio-state.edu!think.com!hsdndev!cmcl2!kramden.acf.nyu.edu!brnstnd
From: brnstnd@kramden.acf.nyu.edu (Dan Bernstein)
Newsgroups: gnu.misc.discuss
Subject: Re: LZW compression patent
Message-ID: <4642:May702:50:2691@kramden.acf.nyu.edu>
Date: 7 May 91 02:50:26 GMT
References: <1991May3.132441.20468@cs.utk.edu>
Organization: IR
Lines: 57