Path: sparky!uunet!mcsun!uknet!axion!srd!psanders
From: psand...@srd.bt.co.uk (Paul Sanders)
Newsgroups: comp.compression,misc.legal.computing
Subject: Talking about patented algorithms in a paper
Message-ID: <1992May19.162240.15237@srd.bt.co.uk>
Date: 19 May 92 16:22:40 GMT
Sender: n...@srd.bt.co.uk (News Administrator)
Reply-To: psand...@srd.bt.co.uk
Organization: British Telecom
Lines: 20

Hi,

I hope someone out there can help me:

As part of a DPhil I have implemented the LZW compression algorithm in a 
parallel functional language and now I want to publish the results. I see
from the compression FAQ that IBM, Sperry Research has patented the algorithm.
(I'm not sure if this is just in the US though).

What ramifications does this have for my paper ? Must I acknowledge the patent,
the author or something else ? I don't know anything about legal stuff so any
advice would be of use. E-mail replies would be best.

Thanks

Paul.
--
Paul Sanders      psand...@srd.bt.co.uk		Tel: +44 473 645716
Systems Research Division, BT Laboratories, Martlesham Heath IPSWICH IP5 7RE
"The sheep was lost; it had fallen off the cragside and into the abyss"

Newsgroups: comp.compression,misc.legal.computing
Path: sparky!uunet!caen!uwm.edu!csd4.csd.uwm.edu!markh
From: ma...@csd4.csd.uwm.edu (Mark William Hopkins)
Subject: Re: Talking about patented algorithms in a paper
Message-ID: <1992May19.194718.8728@uwm.edu>
Sender: n...@uwm.edu (USENET News System)
Organization: Computing Services Division, University of Wisconsin - Milwaukee
References: <1992May19.162240.15237@srd.bt.co.uk>
Date: Tue, 19 May 1992 19:47:18 GMT
Lines: 13

In article <1992May19.162240.15...@srd.bt.co.uk> psand...@srd.bt.co.uk writes:
>Hi,
>
>I hope someone out there can help me:
>
>As part of a DPhil I have implemented the LZW compression algorithm in a 
>parallel functional language and now I want to publish the results. I see
>from the compression FAQ that IBM, Sperry Research has patented the algorithm.
>(I'm not sure if this is just in the US though).

LZW is not patented.  Rather, what IS patented is the application of LZW
to modem transmissions.  Algorithms and other mathematical objects can't be
patented.

Path: sparky!uunet!mcsun!corton!chorus!nocturne.chorus.fr!jloup
From: jl...@nocturne.chorus.fr (Jean-loup Gailly)
Newsgroups: comp.compression,misc.legal.computing
Subject: Re: Talking about patented algorithms in a paper
Message-ID: <13568@chorus.fr>
Date: 20 May 92 08:59:44 GMT
References: <1992May19.162240.15237@srd.bt.co.uk> <1992May19.194718.8728@uwm.edu>
Sender: jl...@chorus.fr
Reply-To: jl...@nocturne.chorus.fr (Jean-loup Gailly)
Followup-To: misc.legal.computing
Organization: Chorus systemes, Saint Quentin en Yvelines, France
Lines: 32

ma...@csd4.csd.uwm.edu (Mark William Hopkins) writes:

> In article <1992May19.162240.15...@srd.bt.co.uk> psand...@srd.bt.co.uk
> writes:
>> As part of a DPhil I have implemented the LZW compression algorithm in a 
>> parallel functional language and now I want to publish the results. I see
>> from the compression FAQ that IBM, Sperry Research has patented the
>> algorithm. (I'm not sure if this is just in the US though).

> LZW is not patented.  Rather, what IS patented is the application of LZW
> to modem transmissions.  Algorithms and other mathematical objects can't be
> patented.

Please read the article in comp.patents:

   From: r...@gnu.ai.mit.edu (Richard Stallman)
   Subject: [SPAT] Compression Patent
   Message-ID: <4...@cluster.cs.su.oz.au>
   Date: 15 May 92 11:11:00 GMT

   [...]
   When the Patent and Trademark Office says, "We do not issue patents on
   algorithms", they are playing a word game [...]

I won't quote the full article, you should read it. It contains the
claims of the IBM and Unisys patents. They are *not* restricted to
modem transmission. The claims are so general that they can cover
compression from disk to disk.

Followups to misc.legal.computing.

Jean-loup Gailly

Path: sparky!uunet!cis.ohio-state.edu!ucbvax!virtualnews.nyu.edu!brnstnd
From: brns...@nyu.edu (Dan Bernstein)
Newsgroups: comp.compression,misc.legal.computing
Subject: Re: Talking about patented algorithms in a paper
Message-ID: <16030.May2102.42.2292@virtualnews.nyu.edu>
Date: 21 May 92 02:42:22 GMT
References: <1992May19.162240.15237@srd.bt.co.uk> <1992May19.194718.8728@uwm.edu>
Organization: IR
Lines: 18

In article <1992May19.194718.8...@uwm.edu> ma...@csd4.csd.uwm.edu 
(Mark William Hopkins) writes:
> LZW is not patented.  Rather, what IS patented is the application of LZW
> to modem transmissions.

False. There's no excuse for anyone on the Internet to mis-summarize the
Welch-Unisys LZW patent, as it's available from ftp.uu.net:doc/lzw-pat*
in scanned form. The patent covers _all_ uses of the LZW algorithm. So
does the Miller-Wegman-IBM LZW/MW patent, which was filed earlier.

As I recall, James Woods spoke with someone at Unisys (Sperry) some time
back, and was told that Unisys only cared about hardware uses of LZW.
This then spread around the net. More recently some rumors have cropped
up of Sperry worrying about software uses. Can anyone substantiate this?
Since IBM's patent counts as prior art it's exceedingly unlikely that
Sperry could win in court. Nevertheless they can certainly threaten
suits. Does anyone know if they have ever done so?

---Dan

Path: sparky!uunet!mcsun!corton!chorus!nocturne.chorus.fr!jloup
From: jl...@chorus.fr (Jean-loup Gailly)
Newsgroups: comp.compression,misc.legal.computing
Subject: Re: Talking about patented algorithms in a paper
Message-ID: <13577@chorus.fr>
Date: 21 May 92 13:57:47 GMT
References: <1992May19.162240.15237@srd.bt.co.uk> <1992May19.194718.8728@uwm.edu> 
<16030.May2102.42.2292@virtualnews.nyu.edu>
Sender: jl...@chorus.fr
Reply-To: jl...@chorus.fr (Jean-loup Gailly)
Followup-To: misc.legal.computing
Organization: Chorus systemes, Saint Quentin en Yvelines, France
Lines: 91

brns...@nyu.edu (Dan Bernstein) writes:

> As I recall, James Woods spoke with someone at Unisys (Sperry) some time
> back, and was told that Unisys only cared about hardware uses of LZW.
> This then spread around the net. More recently some rumors have cropped
> up of Sperry worrying about software uses. Can anyone substantiate this?

Here are extracts from two articles which appeared earlier in
comp.compression and comp.patents. The usage of LZW in pkzip 1.10 is
definitely a software implementation. The usages in V42.bis and postscript
(both licensed to Unisys) also appear to apply to software implementations
(see extract below).

Jean-loup Gailly
jl...@chorus.fr

  ---------------------------------------------------------------------------
  From: sid...@borland.com (Sidney Markowitz)
  Newsgroups: comp.compression
  Subject: Re: PKZip 1.93a performance
  Message-ID: <1991Dec19.011750.5...@borland.com>
  Date: 19 Dec 91 01:17:50 GMT
  Organization: Borland International

  [...]
  Given prior discussion in which people claimed that Unisys is
  enforcing their LZW patent only for hardware implementations, I
  thought that the following messages I saw on Compuserve would be of
  interest:

  [begin quote]

   #: 133873 S0/Gen/New Uploads [S]
      06-Dec-91  16:08:48
   Sb: #132994-#PKZIP Alpha
   Fm: Tom Palmer 76244,3564
   To: PKWARE - Doug Hay 75300,730 (X)

   It's too bad that "-es" has slowed down so much in the new alpha.
   Perhaps not in the final version -- ?

   #: 134092 S0/Gen/New Uploads [S]
      09-Dec-91  09:37:36
   Sb: #133873-#PKZIP Alpha
   Fm: PKWARE - Doug Hay 75300,730
   To: Tom Palmer 76244,3564 (X)

   The -es no longer uses LZW compression.  UNISYS wanted VERY
   unreasonable amounts of money to use the algorithm.

  [end quote]

   -- sidney markowitz <sid...@borland.com>

  ---------------------------------------------------------------------------
  From: edsr!tantalum!che...@uunet.uu.net (Mark Costlow)
  Newsgroups: comp.patents
  Subject: Re: Permission to use compress
  Message-ID: <1992Jan3.025034.4...@edsr.eds.com>
  Date: 3 Jan 92 02:50:34 GMT
  Organization: EDS Research, Albuquerque, NM

  brns...@KRAMDEN.ACF.NYU.EDU(Dan Bernstein) writes:

  >Alan Deehr writes:
  >> Can anyone provide any concrete information on the legal aspects
  >> of using the public domain versions of the source code for
  >> compress?
  >
  >Sure. There are no copyright problems with compress. However, patents
  >4,558,302 and 4,814,746 appear to cover LZW (aka MW1) coding, which
  >compress uses. Since 4,558,302 (Unisys) was filed after 4,814,746 (IBM),
  >it stands almost no chance of holding up in court. However, according to
  >a statement apparently from a PKware employee, Unisys has been quietly
  >extorting (uh, I mean, asking) license fees for 4,558,302.

  I recently saw the following in "PostScript(R) Language Reference Manual,
  Second Edition", by Adobe Systems Incorporated:

	  [PostScript Level 2 implements several compression filters
	  in the interpreter itself.]

	  "The LZW compression method is said to be the subject of United
	  States patent number 4,558,302 and corresponding foreign patents
	  owned by the Unisys Corporation.  Adobe Systems has licensed this
	  patent for use in its products.  Independent software vendors
          (ISVs) may be required to license this patent in develop software
                                                                   ^^^^^^^^
	  using the LZW method to compress PostScript language programs or
	  data for use with Adobe products.  Unisys has agreed that ISVs may
	  obtain such a license for a modest one-time fee."