[lug] ssh compression

[Paul Walmsley]

On Wed, 1 May 2002, Bear Giles wrote:

> > two somewhat pathological examples follow.
>
> I agree - those are pathological examples!  You can only really
> compare compression performance with the type of data that you'll
> actually be compressing.  The pathological cases may be as misleading
> as saying that "trees don't give you any benefit over linked lists"
> when you only test them with data that's already sorted.

the examples are not as irrelevant as they might seem.  my experience has
been that with those two general compression methods, in terms of
compression time, most files will fall somewhere in between those two
poles.  you can verify this empirically, if you like.  i certainly agree
that, ideally, one should test with the data involved with one's own
application.

i hope that others are spared the surprise that i had, when a backup
process of mine that previously took three hours ended up taking over
eight hours when we used bzip2 rather than gzip.  if you take the linux
kernel, for example, compared with gzip, bzip2 achieves a 20% size
reduction in exchange for a 400% increase in compression time.  caveat
user.


- Paul