> Out of curiosity, though, I tried your test case. Lo and behold, Win XP's
> built in compression utility (a ZIP algorithm) did not recognize the
> sequence! I generated a file that contained 65535 16-bit numbers, in
> sequence from 0 to 65534, represented in little endian format in the file.
> It compressed from 127 K to 119K. I don't know whether to blame Microsoft
> or if most compression programs would do the same. I suspect that very few
> compression programs would catch this sequence since it would be rare to
> have such things in real data and it would probably take a special check
> for this specific case in order to detect it in a finite data set.

Now create a file that contains 10 copies of this sequence and try that...

Wouter

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.