From: Saltwell, Erik (esaltwell_at_[hidden])
Date: 2000-10-20 08:46:02
It might be worth treating different compression algorithms here like the
different random number generation algorithms in the boost libraries. By
treating the actual compression mechanism as an external entity that
supports a particular interface, people who were allowed to use these
algorithms could, without having to account for them up-front.
From: Gary Powell [mailto:Gary.Powell_at_[hidden]]
Sent: Friday, October 20, 2000 2:58 AM
Subject: RE: [boost] compression filters
> 1. what would be the relevance of a few generic compression filters to
I'd use them! I've coded up Huffman at a previous job (Aldus/Adobe) for
their installer. (I don't think they use it anymore.) The quick look at the
interface leaves one question, and that is to get optimum compression,
keeping a single global table is often used. I'd rather it was a parameter
to your algorithms. In the game business we often want to encode chunks of
data within a file. (A particular animation or texture) The file is full of
similar data so the encoding table can be reused. The output needs to know
the new size/length of the stream so that we can update the map/table for
lookup in the file. On decode we compare expected lengths to actual to help
keep check file integetry.
The other issue is some compression is copyrighted. So those filters would
be not available except as a sub license.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk