Yawn

I prefer PowerArchiver's .cab personally...but different compression method's are better for different things. I use rar & .cab for my backup's.... .cab beat's out .rar mostly but in some cases rar wins.

Is that a "Normal" compression or a "Maximum" compression? I always found CAB files give pretty crappy compression ratios.
The best I think I ever found was LHA/LZH, which is (or was) used largely in Nihon [Japan]. ACE generally works better than RAR, and bzip2 generally works better than most of them (however, I haven't done many tests with ACE because I don't actually HAVE WinACE -- one compression program is enough for me, thanks). I can't vouch for 7zip. However, LZMA sounds like it uses the same (or simlar) compression technique as LHA/LZH (Lempel/Ziv/Huffman) -- most likely Lemphel/Ziv plus some other dude and "archive". In any case, it seems to me that the best compression could be achieved by creating a format that incorporated many different aspects -- lossless images => PNG, images acceptible to be lossy => JPG, animations => MNG or AVI -- basically just getting rid of all of the uncompressed crap (Targa, PCX, Bitmap). Similarily, using LAME --r3mix -q0 for audio would work wonders -- it's preceptually lossless and still does a damn good job of it. Vorbis would be even better, but it hasn't been fully tested yet so I can't really recommend it. Yet. Anything else...text files, other strange data resources (don't ask me what, since computers can output three things: visual, aural, and more recently "touch" data) could be compressed using whichever format worked the best for compressing, y'know, text or whatever. But then, that's just a dream of mine, and one that's already been at least partially accomplished with MP3 packing utilities from CLASS and other cracking groups, but one that will never really be a reality. (Said CLASS utility would be useful if it was used legally and if it was openly modifiable and if Fraunhofer didn't want to charge $0.75 per decoder now.)
Anyway, whee.

My personal solution to most things like this is to make less bulky shit. 😄 [Vector data, well-written code *cough,ahhem*, no !@#$ raster data]

Oh, and SHA-1? LET'S GET THOSE DISTRIBUTED.NET BOXEN ON IT! Honestly though, CRC32 and MD5 generally work well enough, since all they're being used for is the verification of versions, really, and not for critical, possibly malicious data. I mean anything could be malicious, but, y'know, WHY?

Off-Topic: For the record, I checked through my DOS4GW's last night. I'd forgotten that along the way I'd been updating them with newer versions, to no ill effect. I'm unable to see what really came with the games without reinstalling them, so while I had 10 copies of dos4gw.exe, 6 of them were the same CRC as 1_97_2.exe, and the rest were 1_97.exe. But I think, now, I'll change them all to ver 2.01a. 😀

Bump.

Man these forums are DEAD. 🙁

You say that, but then realise that I'm still getting three or four new users signing up every day. (Not that they post, but...)

Besides, it seems only the regulars or near-regulars pay any attention to anything other than getting their own problems solved.