[cardigan]

Because this attack is not feasible against MD5 with current state of the art cryptanalysis.

Instead collisions between files are demonstrated by appending bytes to two files until their hashes match.

[taneq]

Ah darn, for some reason I thought there was a practical preimage attack on MD5. I must have been thinking of CRC32 (which is just useless for this sort of thing.) My mistake.

[dom0]

CRCs are linear, all 'attacks' are efficient on them.