Re: RDF Update Feeds + URI time travel on HTTP-level

Danny Ayers wrote:
> What Damian said. I keep all my treasures in Subversion, it seems to work.
> 
> 
> 

3rd that; whilst the http time travel conversation goes on - I can't
help feeling that going down the date header route is only going to end
up in something nobody uses; because it doesn't provide any
implementation details to the developer, and thus nobody will adopt it.

subversion/webdav/deltav on the other hand, everybody knows, it already
works, does the trick and would be easy to implement - essentially all
we're saying is let's version control rdf, a concept we can all
understand, and at worst the addition of a http response "version"
header tag would pretty much solve exposing all this functionality
through http/rest etc. We could handle exposing diffs etc via restful
post/get params (?since=r6) and also expose different synchronisation
endpoints for data eg on a graph level or a resource level, or however a
developer chooses to do it; the point is that simply specifying to use
version control and one additional version response header will do the job.

it's not perfect, it's not time travel; but it addresses the need in a
familiar standards based way that's been thoroughly thought through and
tested; and moreover it'll allow us all to get on and sync our RDF, now,
rather than in 2 years when it's too late.

all imho of course.

the only thing I can see that remains is to determine the format /
serialization of the updates, and primarily delete, we can take it for
granted that all normal triples / quads are new - so all we need to do
is find a way of saying X quad / triple has been removed.

kinds regards, and naive as ever,

nathan

Received on Wednesday, 25 November 2009 10:52:16 UTC