In a comment on my recent post about content resolution and the MTV playlist archive, Danny Ayers brings in the semantic web toolkit:

Weaving a XSPFing yarn

I suspect a GRDDL/RDF/SPARQLing hammer could hit this nail on the head too

I'd love to see that done. Though, GRDDL requires XML input, and since the original data isn't in a XML-ish format (yet) it would have to be converted to one, and that XML format would probably be best as XSPF itself. These tools are complementary, not mutually exclusive.

1. source document -> (custom scraper)
2. XSPF -> (GRDDL) ->
3. RDF -> (SPARQL) ->
4. songs

The relationship between SPARQL and content resolution is new to me. It's an interesting similarity. The one reservation I have is that content resolution has an aspect of computational linguistics which would be hard to do in SPARQL. (Given my very limited understanding of SPARQL).

About using GRDDL to convert XSPF to RDF, that's completely in harmony with the design of XSPF -- we always intended it to be an intermediate format facilitating transfer, sharing and repurposing. That it's in the middle of the chain of transforms above is no surprise.