26 October 2004

Genius idea #2

This is an idea I had a while back, but a Web site just reminded me how useful it would be--that's always a good sign for an idea if you keep thinking about it.

There needs to be a Web utility that retrieves and organizes all streaming schedules so that they can be searched. Many radio stations, generally classical, broadcast their playlist or at least a description of thier specialty shows. All of this information could be scraped and imported into a database so that people could search for "Shostakovich" or "Red Hot Chili Peppers" or "symphony" or even "David Foster Wallace interview." Once a reliable schedule database is set up, you could schedule songs or shows to be recorded.

I think of it as Radio TiVo.

This site lists upcoming webcasts of modern classical music pulled from various European radio stations. It may be automatically generated from the source sites, but it looks more like it's built manually (the header points to a source document from someone's Internet cache: C:\Documents%20and%20Settings\Marc%20Adler\Local%20Settings\Temporary%20Internet%20Files\Content.IE5\WPIVSLQ3\2004-04-09%20UpcomingWebcasts[1].htm). I would like to create a normalized database of broadcast media, and develop filters that populate it from different streaming sites on the Internet. A spider would walk around grabbing updated playlists and then pass them to the filter specific for each site.

Although there are many formats and layouts, a generic screen-scraper engine could be written to simplify the task of customization. I wouldn't know the real difficulty until I did some more detailed research. I had written previously about Paul Ford's efforts to screen-scrape government sites. It's very susceptible to breaking whenever layout is modified, but I would hope that some of the more content-rich stations would begin publishing a standard playlist in XML.

It would be a fun project either way.

[ via The Rambler -> Upcoming Webcasts of Modern Classical Music ]

[ posted by sstrader on 26 October 2004 at 11:21:06 AM in Science & Technology ]