Ran some more testing last night after I wrote the second prototype of the compressor (still prototyped in VB, so it’s still slow 😉 This time it to make SURE it worked on really big files – and yep, it does. Took a compressed single-file game setup of 44MB, and reduced it to ~300k. Then I tool an 84 MB video file (something else that doesn’t compress worth a shit) and reduced it to ~340kb. Damn am I ever glad for that dual 1.8 GHZ machine – throwing a lot of horsepower at the situation helped a lot!
I’ll be starting the ‘real’ version of the compressor and decompressor in C++, and a library for programmers wanting to integrate it. This should be an interesting job – I haven’t written anything in C++ from scratch for about three years (was working on a telnet server idea I had) so I’m gonna be damned rusty in doing this for a short bit until I get back into things.
I may talk to RMS (Richard M. Stallman, for those who don’t follow Open Source stuff) at the Free Software Foundation to see if there is a dual-licensed version of BZ2 compression, and see if I can get ahold of the non-GPL or non-LGPL’ed version it integrate directly in as the basis for the compression portion of it (BZ2 is fairly fast. It’s not as good compression wise as some other routines, but it’s also patent free so I don’t have to worry about anyone fryin’ me over patents.) Talking via email to RMS has always been an interesting experience – if you start the conversation off on the wrong foot, he’s an ass. If you start it off right, he’s pretty interesting to talk to. The FSF is one of those groups I both love and hate all at once. *SIGH* Of course, I could just use BZ2 lib, and ignore the FSF – technically, I can use a LGPL library without problems. Or I could just go with something like InfoZip as the base, and chuck it all. Choices, choices.