Free, social hosting, like Flickr, YouTube, but for triples?

Moin all, it seems we are all hosting our own triple stores. I was wondering, though, if there are (or if we will see soon) hosting services for RDF data. Like a SourceForge or GitHub for triples, with a at least a SPARQL end point, but perhaps also a (simple) bug tracker, and certainly tools to create online communities as in Web2.0.

Is anyone aware of any current such (free) hosting services for RDF data?

I just saw Dydra. Is still in private beta, so I haven't tried it yet. It doesn't provide social aspect that you requested, but it provides RDF store, and SPARQL endpoint. I'm also not sure about pricing.

Have a look at the Talis platform. It doesn't provide the social aspects you mention, but is otherwise the closest I am currently aware of. Beyond storing triples, querying and allowing RESTful updates it has changesets with rollbacks.

For qualifying open data sets they provide free hosting to encourage data publishing. However it's free to try it out, and Talis seem like nice people.

Freebase is certainly free, and maybe it qualifies as "social"? They have a notion of per-user datasets and their own query language, query builder etc. Maybe they support SPARQL as well.

I haven't used Freebase much, but their services seem to be biased towards encouraging people to contribute to the open web of data, rather than maintain their own personal data silos. (An admirable goal IMHO.)

I've tried a few and disliked all of them, so I've decided to beat them all and to make my own... and I disliked it too.

The reason is as simple as latency. If I set up a local service then I get instant replies and others suffer from latency. If I set up an endpoint at some remote hosting that is supposed be convenient for everyone then I suffer from latency and it's not much better for other users.

So if a student needs a SPARQL sandbox but does not have an appropriate hardware I give him a login to some of my local boxes. Starting one more Virtuoso is definitely not a problem, and making a symlink from one hard drive to other is the fastest way of "downloading" some initial dataset like DBpedia.

Scientists can have a look at Science 3.0, which is starting free hosting for CC0-waived triples, which is actually powered by the listed Talis hosting.