Comment by pfraze
5 hours ago
Yeah that'd be interesting to try. If you extend the endpoints and the vocabularies around RSS/OPML enough, you could likely create a fairly robust dataset replication protocol. Microformats are a decent schema basis, but you might want to build upon them as well. If you want to make the users' sites reusable as a datastore across applications, you could introduce an oauth flow to enable 3rd party writes. You might also want to add content signatures so you can verify the data authenticity from third parties (e.g. to support the equivalent of reposts).
You would repost a page by linking to it from your own. As for uploading content, that is outside of the scope of an indexer. The user hosts their data with some hosting service, or independently. I believe signatures are already handled by XHTML which allows you to add XML signatures to documents, but regular HTML is sadly lacking here thought it would not be too hard to extend (or to just use the XML serialisation of HTML5).
I think it is generally bad to implement new features like this on the part of the indexer. It should just keep track of an existing web of documents rather than creating it's own format and walled garden.
> You would repost a page by linking to it from your own.
Ah right, of course. Fair enough. I think Dave Winer is trying some similar ideas, though I haven't looked at it beyond knowing he's trying to extend RSS to support these capabilities.