[jwecker]

So by saying 'cannot _always_ work,' and 'very unlikely to work,' you are implicitly conceding the point- that it is possible. The question then becomes, given the size of the OS environment space (and the quality of the randomness in the originals), could one in 50 attempts actually succeed in shaving off a byte? I sincerely doubt that Mike, despite his superior knowledge of information theory, actually took the time to figure out just how good or bad the odds were. He assumed that it was fundamentally impossible, when it is in fact not impossible, just unlikely (to some unknown [to me] degree).

[recursive]

Let me just put it this way. It's more likely that you could win using gzip. It's similar to saying it's "possible" that the target file could have contained all zeroes.

Assuming the target file is competently constructed, the chances of winning wouldn't make it a rational bet for $0.01

[shrughes]

I'm so sorry you had to suffer under these postings.