🚀 go-pugleaf

Another patent to look up, and decypher...

  From: victor@ibm.com (Victor Miller)
  Newsgroups: news.admin
  Subject: Re: Compression Schemes to replace Compress (was: [from comp...)
  Message-ID: <1990Jul30.191350.18977@arnor.uucp>
  Date: 30 Jul 90 19:13:50 GMT
  Reply-To: victor@ibm.com (Victor Miller)
  Organization: IBM T.J. Watson Research Center

  I should point out that Miller (that's me)-Wegman is patented: 
  U.S. patent number 4,814,746, granted in March 1989.  To further muddy
  the waters, some of our claims are the same as Welch's and we filed
  first.  In any case, both patents are dependent on the original
  Lempel-Ziv Patent (Eastman, Cohn, Lempel and Ziv).  Lempel and Ziv have
  described two techniques (I'll call LZ1 and LZ2  -- in order or appearance).
  The patented one is LZ2.  The technique LZ1 was, at first, thought not
  to be practical, but eventually people have adopted variants of it.  The
  most widely used is probably in PKZIP on PC's.  It does have the advantage
  that decompression is very fast and cheap.

  I'm sorry to see that people view M-W as such a hog.  That maybe because
  of the implementation of squeeze.  In general, it should only require about
  130K of working storage for compression.  I've found, that in a good
  implementation,
  it is about a factor of 3 (in terms of CPU time - not wall clock time)
  slower than algorithms like that in compress.  I run it all the time on
  an IBM/PC running DOS with only 640K.


			  Victor S. Miller (victor@ibm.com)
			  itinerant Number Theorist
			  IBM, TJ Watson Research Center