home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.compression
- Path: sparky!uunet!charon.amdahl.com!pacbell.com!decwrl!sun-barr!cs.utexas.edu!qt.cs.utexas.edu!yale.edu!jvnc.net!news.edu.tw!news!nchud5.nchu.edu.tw!NCHUD1.NCHU.EDU.TW!303JKJAN%VAX9K.DNET
- From: 303JKJAN%VAX9K.DNET@NCHUD1.NCHU.EDU.TW (jinn-ke jan)
- Subject: Re: Compression + Standards = Trouble?
- Message-ID: <303JKJAN%VAX9K.DNET.3.720975395@NCHUD1.NCHU.EDU.TW>
- Lines: 50
- Sender: usenet@nchud5.nchu.edu.tw (USENET account)
- Nntp-Posting-Host: 140.120.7.38
- Organization: Dept. of Applied Math. Nat'l Chung Hsing University
- References: <Bwy39p.HxK@news.cso.uiuc.edu>
- Date: Thu, 5 Nov 1992 14:56:35 GMT
-
- In article <Bwy39p.HxK@news.cso.uiuc.edu> jy10033@ehsn20.cen.uiuc.edu (Joshua M Yelon) writes:
- >From: jy10033@ehsn20.cen.uiuc.edu (Joshua M Yelon)
- >Subject: Compression + Standards = Trouble?
- >Date: Fri, 30 Oct 1992 17:29:45 GMT
- >There's been a big discussion under comp.os.linux about which
- >compressor to use... most people think that we should use compress,
- >since everybody has "uncompress", but not everybody has unzip
- >or melt.
- >
- >I was a little bothered, though, that we were tied to an obsolete
- >compression program simply because "uncompress" is the only one
- >available as part of the standard unix distribution. Of course,
- >we could try to convince vendors to start putting unzip or melt
- >into their distributions, but by the time we succeeded, those
- >would be just as obsolete.
- >
- >I therefore started trying to design an obsolesence-proof compression
- >standard. It seems to me that the only way to achieve this is by
- >using self-extracting archives. Any compressor can use any newfangled
- >compression scheme it wants, without fear that somebody won't be able
- >to process the result, as long as the archive is self-extracting.
- >
- >However, self-extracting archives typically have problems of their own:
- >
- > * they include decompression code, which involves overhead,
- > * they are executable binaries, and can transmit viruses,
- > * the included decompression-code is typically machine-dependent.
- >
- >All of these can be solved by writing the decompression code in
- >some machine-independent p-code, with no dangerous instructions
- >(thereby preventing viruses), and a special instruction for doing
- >lempel-ziv decompression (which can be used as a one-line decompression
- >program for those cases when the overhead isn't worthwhile).
- >
- >Unfortunately, a pcode interpreter might be slow, compared to
- >decompression code written in assembly language. Potential solution:
- >don't write a pcode interpreter, write a pcode compiler. Have it
- >cache the compiled versions of decompression headers that it's seen
- >before.
- >
- >What I'd like to know is, what am I forgetting?
- >
- Might be the executable binaries are OK., since the technique of software
- authentication may be used for preventing viruses. For details please refer
- to (1). Okamoto,E and Masumoto, H(1990):ID_Based authentication systems for
- computer virus detection, Electronics Letters, Vol.26, No. 15, July 1990.
- (2). Harn, L. et al(1992): A software authentication systems for the
- prevention of computer viruses, Proceedings of ACM 1992 computer science
- conference, March 1992.
- J. K. Jan.
-