Hah! I just generated as many hashes in a minute that I did in the hour I had it running beforehand. Good changes :)

You might also consider removing UTF8Encoding from SHA1 implementation, I might be wrong, but I think contest states no encoding should be used - i.e. ASCII only. If that's a care, that's a useless O(N) operation.

UTF8(ASCII) == ASCII

Unless you're using characters > 255, UTF-8 is identical to ASCII.

Actually it's UTF8(ASCII) == ASCII iff ASCII < 128, since the MSB signifies a multi-byte character.

http://en.wikipedia.org/wiki/UTF-8

Hm, on second thought UTF8(ASCII) is -always- == ASCII, since ASCII is only defined for 0-127, unless you count the extended variants :P

http://en.wikipedia.org/wiki/ASCII

Yeah, my JS hamming distance implementation is pretty miserable. I'm no bit wizard, but I thikn I can modify the SHA1 function (before it returns the value as a hex string) and XOR the 5 vars with the hex value that they gave us, and then count the set bits. Couldn't get it working last night though so I switched over to this

You could also make use of the fancy new worker thread stuff in browsers that support it. Should let you skip the timeout altogether.

Now we're talking! 7 tabs = 80% cpu usage!

You might also consider lookup tables for Hamming distance (google for "counting bits").