Re: Compression and Unicode

From: Juliusz Chroboczek (jec@dcs.ed.ac.uk)
Date: Thu May 11 2000 - 05:56:36 EDT


Mark Davis <markdavis@ispchannel.com>:

MD> [SCSU] is specifically architected to work well for small,
MD> independent pieces of text.

MD> [...] minimize [runtime] memory requirements

Two good points. Both can be achieved with general-purpose
compression algorithms, slightly adapted. However, having to tweak
the available libraries sort of misses the point of using standard
algorithms in the first place.

MD> [BTW, it was also our experience that with larger files,
MD> compressing with SCSU then compressing with LZW produced better
MD> compression than LZW alone.]

I'm quite willing to believe this. However, I'd be surprised if the
difference were significant (as in: more than a handful of percent
with four fingers chopped off).

                                        J.



This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:21:02 EDT