From MWCSWiki

Contents 1 Big issues to decide 2 S vs SA 2.1 Multi vs single value 2.2 What are concentrating on right now? 3 One possible way to go for divergent data: 4 Big picture thoughts 5 Potential Action Items for 2nd half Spring Semester 2010:

Big issues to decide

S vs SA

Multi vs single value

There are three possible ways to go:

1. The system does not know whether a predicate is single or multi value. It just behaves as if they are all multivalue. (Con: it's weird that the user has to explicitly disagree with a height instead of just changing it.)
2. The system is explicitly told by the user whether a predicate is single or multi value. (when it's created?)
3. The system infers whether a predicate is single or multi value (probably when a second value is added.) (Con: confusing to the user how to distinguish

What are concentrating on right now?

• Import/export
• Divergent data testing

One possible way to go for divergent data:

• An SA interface (for the sake of import if nothing else)
• At creation time, attributes are designated as single or multi value.

Big picture thoughts

• Do OKM users have the same "status" as imported knowledge sources? (SD says yes, CD has heartburn)
• Do we import ontologies/schema, or only data? (SD/CD say both)
• How do we present imported data? (with an HRN. Either we use rdfs:label, dc:title, foaf:name, etc. if they exist, or as a last resort take the last bit of the URI (after the last delimiter.))
  • What about object names? (we could either (a) dereference the object's URI and just get a label, otherwise default to last bit of URI, or (b) just use last bit of URI.)
• How do we deal with duplicate HRNs? (in all search boxes, etc, present the source name nearby if there is a duplication)
• What about exporting data with duplicate HRNs? (we could make our URIs be like this: http://rosemary.umw.edu/okm/okmobject5942#Han_Solo)

Potential Action Items for 2nd half Spring Semester 2010:

1. fix RHR stuff (rolled in by Trillane)
2. general bug fixes from EXP 3[[1]]
3. Bring Divergent data back to the state it was in at the end of SSI
   1. Investigate if any Persistence/Domain changes affect old DD code
   2. Uncomment DD code (including menu-bar (comment out Query still?)
   3. Rinse, repeat 1+2
4. Improve Divergent data
   1. More comprehensive/organized statistics presented on agrees/disagrees