While I appreciate the point made - that any lossy compression of source material is irrevocable - I think it's important to consider the role file-sharing systems play in the spectrum of data storage and retrieval mechanisms.

Now, if resources permitted, the ideal method of preserving and presenting content such as the collection of recordings discussed above would be something along the lines of the Gutenberg project - a set of widely mirrored true-to-the-original files, a searchable index and a website of meta-information. Such a system is very costly to set up and maintain, requring money for bandwidth and hundreds of hours of effort.

At the other end of the scale is the typical fate of realtively upopular resources - a box of unreadable tapes in some enthusiast's attic. Or worse, landfill.

Where p2p file-sharing systems excel (or will do if legal issues caused by copyright abusers stop hampering their popularity) is in providing an option between thse two extremes. Material can be made available online easily and cheaply, albeit with a minimum of metadata and no index. The materials persistance only guaranteed if it is popular - the day the last person sharing it decides that they are no longer willing to donate their disk space and bandwidth to the material it will be gone.

It's a trade-off. I haven't the resources to do the 'right' thing of properly archiving the material, but I'm not willing to simply let it vanish. This trade-off extends to having to take a pragmatic approach to bandwith/quality issues. Yes, it'd be marvellous to not have to permenantly degrade the recordings, but a box of CDs in my attic would be of as little use as a box of tapes. All I can do is make the best of the meagre resources I can spare (here about 3MB/minute) and try to ensure the collections permenance by providing it in popular formats, thereby maximising the chance that others will propagate it.