The podcast directory at ipodder.org should be the center of the podcast world. Key podcast clients like Doppler and ipodder rely on the OPML (an outliner format that is XML based) feed at ipodder.org to offer the full range of podcasts from within their programs. I have experimented with an XSL transformation that can crawl the ipodder OPML and provide some more descriptive information on each of the podcasts. It sort of works (see the results here from November's crawl). I tried the crawl again a few days ago and had terrible problems with a number of special characters from other languages. I was forced to rip out a number of the entries in order to finish the crawl. In addition, many OPML links can still not be crawled due to the encoding problems. (Take a look -- it's not pretty. Technology and Travel nodes still cannot be crawled). Not good; especially when you see that both Doppler and ipodder clients have been crippled by the same problems. As of today, none of the major clients can crawl the full directory.
You know, it's great to claim that OPML can pull together a number of disparate nodes and allow editing of the directory by dozens of people but the problem now is that we have encoding and character-set problems. I think the ipodder directory needs to standardize on a language and a character encoding scheme that is universal (I vote for English and UTF-8). This needs to get done pretty soon or else there will be a further erosion of faith in the ipodder.org directory. Locations like podcastalley.com are ostensibly doing a better job right now. The promise of an open and crawl-able directory is great but it's time to take the Babel out of it.
Add new comment