Reporting Meeting, 2010-01-11

See Reporting Meetings for meeting logistics.

Agenda and Minutes

Agenda

1. Update on outstanding actions
2. Status review
   • schema
   • query
   • REST api
   • service discovery
   • resource format requirement
3. Query - ArthurRyman

Outstanding Action Items

• Action Item: TackTong and DragosCojocari to address the new use case based on RSx in spec. Note: Will be done later.

Minutes

Attendees: SebRose, JamesMoody, XiangDongHu, ScottBosworth, DaveJohnson, ArthurRyman, EinarKarlsen, PaulMcMahan

Regrets: BenjaminWilliams, DragosCojocari, SteveSpeicher

1. Welcome DaveJohnson and EinarKarlsen to the WG.

2. TackTong briefed the overall status of the following topics.

• schema
• query
• REST api
• service discovery
• resource format requirement

3. ScottBosworth pointed out that OSLC-CM may also looking into decriptions of resources ("schema").

Action item: TackTong to sync up with SteveSpeicher on CM's use case on "schema" and align proposal to cover both Reporting and CM use cases.

3. ArthurRyman led discussion on Query proposal.

• In the current proposal, the response to the query includes an rdf:about tag whose value is the subject resource URI with the query parameters appended. This is probably incorrect, or at least inconvenient. It would be more appropriate to omit the query parameters on the URI since then the result can be directly interpreted as RDF statements about the subject resource, and these can be directly added to a graph that contains other information about the subject resource.
  
• Will allow for prefixes to be defined on the URI. Previously, the term "namespace" was being used for these prefixes, however it makes more sense to call them prefixes since that is how RDF formats refer to then and in general they are just the initial segment of a URI and not necessarily namespaces.
  
• The proposal assumes that the query author knows which resources are collections and what their member properties are. This can be determined from the service speciifcation or some machine-readable specification, e.g. the Reporting metadata or RDFS.
• Paging is a cross-domain topic and not part of Query specification.
  • Action Item: TackTong to table paging for discussion in this WG.

4. A subsequent session was held to continue discussion on Query. (attendance : ScottBosworth, DaveJohnson, SebRose, SteveSpeicher, PaulMcMahan, JoanTouzet, BenjaminWilliams, ArthurRyman )

• The workgroup agreed that using RDF as the conceptual data model was OK, but that the more RDF technology being introduced, e.g. SPARQL, the less appeal the specs will have.
• Need to make it clear that RDF is being used to describe the meaning the specs, and do not require them to be implemented using RDF technology.
• RDF/XML is just a format for representing the conceptual data model. JSON, ATOM are other possible format as well.
• There are actually two distinct audiences for the spec, people who write queries and people who implement query engines. Those whose write queries just need good examples. Those who implement query engines may appreciate the SPARQL semantics if they want to test their implementations.
  • Move mentioning of SPARQL into another document for implementors
  • This spec. will be just using good examples

Next week

Comments

Add your comments here:

%CMMENT%