A nice take at Apple's iTunes site, via 2lmc. Also a discussion on bzip2 vs. gzip. The arguments given there, i.e. speed vs. compression rate seem rather academical to me – people are using Stuffit for heaven's sake. Nobody really cares about that. To me the appeal of bzip seems to be some extra geekiness factor plus an additional 'open source seal of approval'.
Of course people will want to download their Linux kernels in bzip. Not for efficiency's sake of course – but because using bzip seems more knowledgeable – idiot savant, if you wish, but savant nonetheless. Like the stage where everybody started using the 'joint stereo' for making mp3 files because of the higher perceived coolness.
If people were really to save bandwidth, say, they could as well turn off their p2p program a few minutes earlier and compress their large downloads with more traditional programs. As my personal Unix god used to point out (in 1993, mind you): transferring an image of your coffee machine is utterly inefficient as all the relevant data easily fits into a single data packet. All a question of priorities, I guess.
Reading the manual pages of both tools is fun as well. That of bzip is a bit less mature. It considers reference to gzip – not directly of course – necessary in the second sentence rather than detailing the way it works first.
Alex, who stayed at our place for the past half year or so, is moving out as she's got a job now. We had a nice dinner tonight and she gave us a cordless phone as a thank you. That's a very nice thing – particularly as our phone is currently deteriorating to a state where it's quite hard to dial numbers... now we'll have to see how much trouble it is recovering the cordless phone from someone's room when you need it. Fun.
Um, isn’t bzip2 about saving bandwidth on the server side? Kernel.org doesn’t pay for itself.
In most cases, whether a given random text file compresses to 82k or 85k is pretty irrelevant, and often not worth the extra compression time (and aggravation because of bzip2’s smaller installed base), but when someone’s shipping big bundles of stuff around, even a 2% difference in size can amount to a significant cost savings. I’m on the Mozilla web site right now, and the source archive in bz2 is over 20% smaller (a little over 8 megs) than the gz version. Multiply that by a few hundred or thousand downloads and it starts to matter, quickly.
Received data seems to be invalid. The wanted file does probably not exist or the guys at last.fm changed something.