Pardon my ignorance, but I could not find an easy way to find the
actual URL of a sparql endpoint, given the URL of the dataset... I
tried querying Sindice over voID descriptions, but it didn't work (or
maybe I'm making the wrong query). In any case, it is not guaranteed
that all datasets will have a voiD description, even if desirable.
Perhaps a simple convention could be adopted by all repositories to use
a "standard" url format, such as <url>/sparql, (e.g.,
http://www.dbpedia.org/sparql), as is being done by many already...
This could provide a simple automatable way to discover endpoints. I
don't mean this as a substitute for voiD, but as an inexpensive