Semantic resources project/CandidateResources/Resources/SWAN Resources/Paolo Questions

First Set of Questions
1. What's the difference between swan:authorSequence and swan:hasAuthorsList, when applied to Journal articles? They both appear to point to the same author lists -- is one of them order-sensitive? Do they both need to be carried over into the new ontology?

2. What about "blank" fields, such as the blank values of swan:hasEvidencetypeQualifier for some of the swan:DiscourseElement entries? These are xsd:string values, but they are given as simply empty strings ("").

3. Do we need to carry over all the new swanadmin:* stuff? Users, passwords, etc?

4. What to do about apparent duplicate KnownPerson entries?

urn:lsid:swan.org:knownperson:db73a10d-fea8-4711-9208-db4e0b50e600 = "J T Finn"

urn:lsid:swan.org:knownperson:add34793-77c0-4065-80ac-672a4795b857 = "John T Finn"

5. Three discourse elements:

urn:lsid:swan.org:comment:4e11bfd7-446d-4b7a-9d7e-61cf500de406 urn:lsid:swan.org:researchstatement:d6e843ca-b83a-4799-b853-3178607f660c urn:lsid:swan.org:comment:d28394df-3285-4011-8c53-636d9c4b8005

appear to be "drafts" -- how should these be handled?

6. What role is "swan:location" playing for Digital Resources? (There appear to be only 16 of them with this attribute.)

7. Why do some owl:DatatypeProperty elements have their domains modeled differently? Contrast swan:journalName, whose domain is an owl:unionOf, with swan:fullName, whose domain is modeled as three separate elements: swan:Gene, swan:Protein, and swan:Organism? Is this an RDF-ism that I'm not familiar with, or are they both representing union types?

8. Three swan:KnownPerson elements appear to have no names (or other identifying information?) at all:

urn:lsid:swan.org:knownperson:74a63b4e-1650-4dea-92bf-fa050d7721f2 urn:lsid:swan.org:knownperson:27815306-27d3-40b8-977e-33e991cdb9c7 urn:lsid:swan.org:knownperson:7c196628-dd7a-4447-b0e5-56695fdd23e4 urn:lsid:swan.org:knownperson:1564d594-1407-449a-aee9-80900df7b95a

All four appear to have been entered by Elizabeth, over four different months (from July 2008 to January 2009), so they don't appear to be incomplete drafts of the same person.

What should I do about these elements?

Second Set of Questions
(1) Do the swan:contains values for DiscourseElements match up with the list structures

(swan:containsDiscourseList -> swancollect:firstItem -> swan:pointsToDiscourseElement)?

If they do match up, is the ordering implicit in the list formulation important?

Are these lists of "contained" discourse elements related to the discourse relationships at all? (In other words, does one discourse element really "contain" another, or does containment mean that one discourse element refutes, supports, discusses, responds to, is motivated by, or is an alternative to another element?)

(2) What's the situation with the swan:researchStatementQualifier (from swan:DigitalResource elements to either core.owl#Claim or core.owl#Hypothesis ... but these are to a different URL, and so they aren't getting prefixed right?)

http://swan.mindinformatics.org/ontology/1.0/20070313/core.owl#Claim

is the target of swan:researchStatementQualifier, but all the other swan: prefixes start with:

http://swan.mindinformatics.org/ontology/1.0/20070313/core.owl#

That looks the same! Is there a character encoding issue with these two URIs?

(3) What about the issue of swan:creationDate and swan:modificationDate?

The documentation in the OWL code for swan:SWANThing seems to imply that every SWANThing should have a creation date (owl:cardinality = 1).

However, there are lots of objects that don't have any creation date at all -- for example:

urn:lsid:swan.org:gene:b704ac10-562f-41bc-8e7f-b6d10475bf7f

which is a swan:Gene. Oddly enough, this element has a *modification date* and timestamp --

swan:modificationDate "2008-07-31T10:51:29.687Z" swan:modificationTimestamp "1217501489687"

Which is doubly-weird, since I don't have a creation date for it.

In general, the question of handling creation dates for the SWAN conversion seems open to me -- should the conversion count as a new "creation," or should this be a "modification?" Or do creation/modification events fall under the rubric of administrative elements which shouldn't be converted?

(4) SWAN 1.0 appears to contain six different discourse relationships. Three of those relationships (inResponseTo, motivatedBy, alternativeTo) seem to have corresponding relationships in SWAN 1.2.

However, the other three (swan:refutes, swan:supports, and swan:discusses) don't have corresponding properties, and it's not clear (to me) how to translate them -- or whether they should even be translated one-to-one.

(5) SWAN 1.0 has three types of DiscourseElements: Research Statements, Research Questions, and Comments. SWAN 1.2 includes a new, fourth kind of element: the "Canonical Research Statement." Should this class be populated in any way during the translation?

What is a canonical research statement, and how does it differ from a "normal" Research Statement?