From DBWiki
It seems that this format would be a bit complicated for a media info parser. Getting a song title requires parsing "span" elements and their class and extracting the child text, getting a preview image requires parsing "img" elements and their class, and getting the media URL's requires parsing anchor elements and their "rel" properties. It would be nice to have a single property type (e.g., "hmedia") in all elements to indicate that media info is present and its type.
Also, the media "type" info uses the displayed text to specify the type (e.g., "song"). However, on a Spanish site, the displayed type might be "aria", but a media info parser shouldn't have to understand Spanish in order to determine that the media is a song.
