Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!elroy.jpl.nasa.gov!jarthur!nntp-server.caltech.edu!madler
From: madler@nntp-server.caltech.edu (Mark Adler)
Newsgroups: comp.compression
Subject: Re: Modeling vs encoding (Re: Lempel-Ziv v/s huffman encoding)
Message-ID: <1991Apr2.102623.13147@nntp-server.caltech.edu>
Date: 2 Apr 91 10:26:23 GMT
References: <1991Apr01.053616.3665@looking.on.ca> <1991Apr1.191751.4211@nntp-server.caltech.edu> <1991Apr02.055023.27834@looking.on.ca>
Organization: California Institute of Technology, Pasadena
Lines: 20


Brad Templeton wonders about the use of fixed trees in PKZIP:

>> If PKZIP does do that, I am at a loss to explain why it sends the full
>> tree in the file

I think the intention was to do two passes, but it was too slow.  It was
probably left in to allow for a two pass compressor in a later version.

>> but more to the point, why it builds a complete temp file for
>> the file and does a full two pass compress?

Does it?

Anyway, I just checked it myself and verified that there are only three
sets of trees in a large number (~100) of imploded entries produced by
PKZIP 1.10.  I can even send you the trees, if you're curious.

Mark Adler
madler@pooh.caltech.edu