4 ms·
The Growth of Linked Data
- thomasfl 16y agoLots of interesting linked data to consume and use in mashups here.
- cschneid 16y agoI looked into RDF pretty heavily a few years ago. The biggest takeaway is that it was naive in its level of trust. If I make a new RDF document in a repo, claim it was written by Shakespeare, and then somebody does the search on "Complete works of Shakespeare", I'd show up. And there was no real way to police that. And unfortunately, we get into web of trust type issues, which in itself is ugly. I really do wish that RDF was usable for search engines, and google, and that they could trust sites. But it's just not feasible, and I don't see any scalable solution to it.
- th0ma5 16y agoI understand this from a standpoint of just general trust, but I don't see it as a problem from technological trust. Had someone else done this, and you came about it, you would probably never go to that site again. The trust layers in the SemWeb stack are just like HTTPS, but anything beyond that is a valid philosophical argument outlined by Cory Doctorow's Metacrap piece. Arguably this is also the open world vs. closed world problem, too, and the concept of authority is purely a technical one.
- cschneid 16y agoFrom a pragmatist standpoint, something like google can't trust an RDF document any more than it can trust the META keywords tag. Anything a site can assert about itself will end up as fodder for spam and manipulation. Not to say that individual trusted datasets don't benefit from the RDF logic layers you can apply. Just that it's mostly impossible to have anything more than a set of curated datasets.