Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!batcomputer!cornell!rochester!pt.cs.cmu.edu!o.gp.cs.cmu.edu!andrew.cmu.edu!vd09+ From: vd09+@andrew.cmu.edu (Vincent M. Del Vecchio) Newsgroups: comp.compression Subject: Re: Number theory compression ? Message-ID: Date: 10 Apr 91 01:53:35 GMT References: <1991Apr4.150053.29873@linus.mitre.org> <1991Apr5.064220.18509@dde.dk> <1991Apr6.080615.23197@zorch.SF-Bay.ORG> <28185:Apr700:26:0491@kramden.acf.nyu.edu> <1991Apr7.020617.12261@nntp-server.caltech.edu> <5351@ns-mx.uiowa.edu>, <4931@pink1.UUCP> Organization: Carnegie Mellon, Pittsburgh, PA Lines: 13 In-Reply-To: <4931@pink1.UUCP> Considering that each input sequence has to yield a unique output, it seems to me that the amount of data that you would need to have to specify where within pi or sqrt(2) your sequence occurs would probably be fairly large. In fact, I don't think it would necessarily be any smaller than the input sequence, unless you could find a sort of "region" of pi which contains many sequences that people are interested in compressing, and sort of prioritize various regions based on how desirable it is to many the sequences in that region compressible.... Sounds like a sort of massive Huffman encoding, doesn't it? Anyway, it's an interesting idea, though I imagine it wouldn't work very well in practice, at least not without a lot of thought. -Vincent Del Vecchio vd09@andrew.cmu.edu