The “combo” design is: an authoring tool, a player / reader, a data format, and an information structure. So, in this case of web music:
* a way to publish song web pages
* a way to play song web pages
* a song web page format
* an overarching structure for what songs are in relationship to things like albums and artists, and to each other
At this point, since we’re talking about info in HTML / web pages, I think the real hurdle is getting a media player that is primarily web page oriented, rather than binary file oriented. If we could get some players to treat the web page as the “song” instead of treating the media file as the song, that would catalyze a lot of efforts around authoring, and exploring options with the format and structure.
So, in this particular chicken / egg problem, I definitely think it’d be good to get some chickens first!