Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!rpi!batcomputer!cornell!rochester!pt.cs.cmu.edu!g.gp.cs.cmu.edu!tgl
From: tgl@g.gp.cs.cmu.edu (Tom Lane)
Newsgroups: comp.compression
Subject: Re: Modeling vs encoding (Re: Lempel-Ziv v/s huffman encoding)
Message-ID: <12564@pt.cs.cmu.edu>
Date: 1 Apr 91 19:24:56 GMT
References: <12546@pt.cs.cmu.edu> <5173@ns-mx.uiowa.edu>
Organization: Carnegie-Mellon University, CS/RI
Lines: 23

In article <5173@ns-mx.uiowa.edu>, jones@pyrite.cs.uiowa.edu (Douglas W.
Jones,201H MLH,3193350740,3193382879) writes:
>
> > A fully adaptive model starts from innocuous assumptions (say, all symbols
> > equally probable).
> 
> Since the compresser and the expander must be primed with identical
> assumptions about the source alphabet, the initial assumptions may
> be used as a key for encryption.

Check.  Another advantage of fully adaptive models for encryption purposes
is that any wrong guess as to a decoded symbol will trash the remaining
output (since the decoder's model will now be out of sync).  This makes the
would-be decryptor's job much harder.

The downside of this, of course, is that adaptive decompressors have
essentially no recovery ability after errors in the compressed data stream.
Fixed-model Huffman decompressors can often resynchronize within a few
symbols after an error.

-- 
			tom lane
Internet: tgl@cs.cmu.edu	BITNET: tgl%cs.cmu.edu@cmuccvma