Time/Place

This meeting is a hybrid teleconference and IRC chat. Anyone is welcome to join...here's the info:

Attendees 

Agenda

  1. ACL discussions from HydraConnect
    1. See: https://gist.github.com/escowles/69df7f2fed1567d83a1b and http://www.w3.org/wiki/WebAccessControl
  2. Moving towards 4.0 release

Minutes

  • ACL in HydraConnect
    • Esme: Hydra wants to have standard Fedora 4 ACL system, would like feedback from Islandora/other users
      • Two approaches: one is like Fedora ACL & Hydra schema, other is emerging W3C spec, but not fully baked yet
    • Aaron: different approaches: predicates on objects declaring access, or external objects referring to objects and policies
  • Nikhil: Database connector
    • Modeshape classes aren't readily extendable, talking to Modeshape developers to make that easier
    • This would be Mode 4.1, so Fedora 4.1 or later
  • Stefano: Mike brought up ordered lists - want to be able to order child nodes in some cases
    • Considering MPTT modified-preorder-tree-traversal
    • Esme: In Hydra RDF, we've used rdf:List, and talked about the ordered list ontology (which seems like a nicer solution)
    • Greg: Ordering harder to validate the structure
    • Mike: maybe we could convert these to ordered nodes in Modeshape?
    • Esme: this common use case, so it could be a good feature, but it could have performance implications
    • Greg: are they fixing the modeshape performance for many children?
    • Esme: they are working on it, not sure if it is fixed or if it can be
  • Stefano: internal sparql query: didn't seem to work, what's the status of this?
    • Longshou: can only query properties that are persisted in Modeshape
    • Stefano: want to have real-time access, access to stuff that's not indexed.

6 Comments

  1. Unknown User (escowles@ucsd.edu): I agree that Ordered List Ontologies are more straightforward to implement and probably more efficient then MPTT/Nested Sets. How do we go about using blank nodes in a Fedora OLO? In the examples provided, the predicate of a olo:slot is a BNode. 

    1. Unknown User (escowles@ucsd.edu)

      You should be able to use blank nodes – they will get Skolemized into URIs that start with /.genid/well-known/.  Though if you want to avoid blank nodes completely, I think you could just create child nodes of the object in question instead.

      1. I raised the question about BNodes a couple of committer meetings ago. See my comment: Re: 2014-10-02 - Fedora Committer Meeting

         

      2. Basically my problem is that I don't know how to relate the BNode to the original triple it was constructed in. 

        1. Unknown User (escowles@ucsd.edu)

          If you PUT that same RDF, the link to the blank node is created, but when you use SPARQL Update, some internal identifier is used as a string for some reason.  I've filed a bug for this: https://www.pivotaltracker.com/story/show/80483482