Doug Kaye's post about Audio Excerpt Blogging is great reading and put a smile on my face. If it is possible to grab a subpart of a mp3 file using a RESTful interface, i would be very happy.
However, something hits odd about using all this other methods to do it (flash, quicktime, etc). Why not just use Smil to do this? It has features to deal with snips of audio already. Plus it has features which you guys will be pleading for soon. I mean who wants to listen to a 2min piece at 56k when there on a T1 connection. Using the switch element you can draw from different sources based on connection speed and other things. Hell, you can even use the same method to take a subpart of a video file… Smil is the way forward for sure, trust me on this one gillmor gang.
However Jon Udells post about Page addressable PDF is pretty cool and I'll be using that more from now on.