Re: Re: Any news on Compress::LZW?

[diotalevi]

I'd have done this myself if I didn't already know that there might be good ways to cheat and get better performance on memory, cpu or on the API. I really wanted SomeGuru to pick it up. *shrug* Having the original paper around would be pretty handy as well.

[meredith]

SomeGuru, eh? Don't know him. :) Actually, I looked at the code (dont tell:) and it's just the LZW algorithm in perl... it doesn't output in any sort of data format, so there's work to be done there, and it doesn't have a decompressor (likely because there is no set data format :). I'm going to poke around some more on the subject... I'll also look at that Compress::SelfExtracting to see if the LZW can be isolated (and optimized, if it really is dog-slow).

Update: I've pulled the compression code out of Compress::SelfExtracting. It's working perfectly... got a 2.91:1 compression on a text file I had laying around :) We'll see if the author cares about his code being repackaged :)

mhoward - at - hattmoward.org