From: Steve Summit (scs@eskimo.com)
Date: Wed Sep 20 2006 - 21:03:51 CDT
Mark Davis wrote:
> I strongly suspect that all of that would give only minor advantages
> over general-purpose algorithms like ZIP.
I have to agree. A lesson which I think has been amply shown is
that the general-purpose compression algorithms are very good.
They can, in fact, "discover" most of the regularities in data
which you might have hoped a special-purpose algorithm could more
optimally exploit. (Discovering regularities is, after, all,
exactly what a general-purpose compression algorithm *does*.)
The engineering expense of designing, deploying, maintaining,
and interoperating a special-purpose compression algorithm is,
generally, not worth it.
> But this is all academic -- I don't see anyone taking the time
> and effort to investigate it in the absence of a compelling need.
It would be pretty easy to do a small but convincing demonstration,
I think -- I've half a mind to do it right now.
This archive was generated by hypermail 2.1.5 : Wed Sep 20 2006 - 21:09:17 CDT