From W3C Wiki
Jump to: navigation, search
  • find a permanent home for AtomOwl
  • the documentation is not quite done (I am hoping Danny will put his excellent writing skills to help out there)
  • improove python script to extract more information out of the N3 ontology
  • It would be good to have a couple of implementations to see if things really work out. This will probably help bring to light some issues. The current BlogEd works on an ontology that is something like this one, but much more primitive. So it should work out ok.
  • It would also be good to see how well what we have meshes with the recent threading extension
  • A few more links were added to the iana registry, these need to be dealt with
  • Seeing how this fits in with Atom APP will also point out problems or tweaks
  • It would be good perhaps also to have an xslt transform that produces output in the form of a tree such as AtomRDF.

<,2005/feedId1>  a :[[FeedId]];
         :update [ a :Entry;
                   :title "my first entry"^:text;
                   :updated "2005-..."^^xsd:dateTime;
                   :id <,2005/entry1>
         :update [ a :Entry;
                   :title "my first entry improoved"^:text;
                   :updated "2006-..."^^xsd:dateTime;
                   :id <,2005/entry1>
                 ] . 

As I think that is the structure that many RDF aggregators would in the end be most interested in. This is a really interesting topic btw, and I am not quite satisfied the notion of a Feed has been quite nailed down. Something in my mind still has to go "tilt". Let me send another mail on this topic though so as to keep this thread on topic.

  • tests: A year ago when I (bblfish) saw Danny in Italy, he had the idea of putting together a number of tests to translate feeds into awol and then do some consistency checks on the results. Other test ideas would be welcome. EliasT is working on a java library to read atom xml into object graphs, which could then be used to fill up a triple database, whose contents could then be compared with the results of xslt transforms.
  • relations to other well know vocabularies such as foaf, sioc, dc. A few such vocabularies are allready used. Should other such vocabulary mapping be placed in external files or be part of the ontology directly?
  • mapping to RSS1.x ontolgies. (of the form ` rss1:Item owl:sameAs awol:Entry ` if possible). It will help to see if Atom is in any way an improovement over what came before itn, and in what way. (bblfish: I have no opinion on this issue and don't mind to discover that not much has been added. Finding this out would be an interesting discovery. Though it is a discovery that would at least be made possible by the ontology (so my time won't have completely been wasted))
  • SPARQL readiness. One of the important aims of this ontology is to make it natural for some one who knows rfc 4287 to understand how to query a SPARQL end point. Of course it will be interesting to hear from SPARQL experts if good querying capabilities have been achieved.