The existing RDF use case

In recent conversations it has been argued by Steve Speicher and Roger Menday
that the ldp:DirectContainer and the ldp:IndirectContainer
fullfill a use case that is to allow smooth transition from sites publishing 
RDF resources to LDP. (I am not sure if that use case was ever documented )

I don't believe that this was ever looked at in detail. I mean have we ever
considered real published RDF on the well known linked data sites to see
if the solutions we are offering are really going to be that helpful?
It may be that alternative solutions would work better. Or perhaps we are
only covering a small set of use cases and the problems may lie elsewhere.

The idea seems to be that these sites that just publish RDF now, may just 
need to add a few ldp:container* relations to their RDF and presto these 
resources can be turn into containers. Old clients using the currently 
published LinkedData could then continue to  work, but new LDP clients would 
be able to infer the ldp:contains relations from existing data in the LDPC 
and could work out which information resources were such that by DELETing 
them one would remove the relations from the LDPC.

Could we perhaps look at some such published sites data and see if things
really work out? What problems are raised when people try to do this?
Can we put forward a Good Practices guide for when people make such a 
transition?

I can imagine problems such as trying to work out on such containers
what triples can be PATCHed and which cannot since some are owned by
the container (unless one allows deletion of ldp:contains resources 
through PATCHing I suppose. )

I am not saying this is not the right solution. It would be nice if
there were some good theoretical reason why it must be the right solution.

Henry

Social Web Architect
http://bblfish.net/

Received on Sunday, 15 December 2013 18:23:41 UTC