W3C home > Mailing lists > Public > public-swbp-wg@w3.org > November 2005

RE: SKOS Core 2nd review

From: Booth, David (HP Software - Boston) <dbooth@hp.com>
Date: Sat, 5 Nov 2005 13:45:51 -0500
Message-ID: <A5EEF5A4F0F0FD4DBA33093A0B07559008911C53@tayexc18.americas.cpqcorp.net>
To: "Miles, AJ (Alistair)" <A.J.Miles@rl.ac.uk>
Cc: <public-swbp-wg@w3.org>

Miles,

Here are my comments on the SKOS Core Guide,
http://www.w3.org/TR/2005/WD-swbp-skos-core-guide-20051102

This is really good work.  Excellent for the SWBP WG to produce.  Kudos!

Suggestions:

1. The Henry VIII example seems to conflate the concept of Henry VIII
with the definition of that concept.  Some triples pertain to the
concept of Henry VIII, others pertain to its definition.  For example,
in these triples:

	http://www.example.com/concepts#henry8
		rdf:type
			skos:Concept .

	http://www.example.com/concepts#henry8
		skos:prefLabel
			"King Henry VIII" .

	http://www.example.com/concepts#henry8
		dct:modified
			"2005-02-06" .

the triples involving rdf:type and skos:prefLabel are intending to refer
to the concept of Henry VIII, but the triple involving dct:modified
pertains to the definition of that concept (as does dc:creator).

To fix this, you could create another URI to identify the concept of
Henry VII:

	http://t-d-b.org?http://www.example.com/concepts#henry8

and leave http://www.example.com/concepts#henry8 to identify the
definition of this concept.  These two URIs would be related by the
triple:

	http://t-d-b.org?http://www.example.com/concepts#henry8
		skos:subjectIndicator
			http://www.example.com/concepts#henry8 .

so the triples above involving rdf:type, skos:prefLabel and dct:modified
would become:

	http://t-d-b.org?http://www.example.com/concepts#henry8
		rdf:type
			skos:Concept .

	http://t-d-b.org?http://www.example.com/concepts#henry8
		skos:prefLabel
			"King Henry VIII" .

	http://www.example.com/concepts#henry8
		dct:modified
			"2005-02-06" .

2. It might be good to mention how skos:subjectIndicator differs from
rdfs:isDefinedBy.

3. Regarding skos:Collection, the "milk by source animal" is good in
terms of being understandable, but I come away from it thinking "Are
collections really needed?  Why not just usea superclass?"  It might be
helpful to mention that even though some collections could alternatively
be modeled using a superclass instead of a skos:Collection, there are
other examples for which there is no clear, appropriate superclass.

4. Regarding skos:ConceptScheme, it would be helpful to use a more real
world example, with a real-ish sounding thesaurus name instead of
http://www.example.com/conceptscheme .

5. ALthough skos:primarySubject seems useful, why limit it to one value?
Would there be a problem in suggesting that it have one value, but
permitting it to have more?  What is the skos:primarySubject of Romeo &
Juliet?

6. Similarly, does skos:prefLabel REALLY need to be limited to a single
value?  Certainly it is most useful if it is unique, and that should be
encouraged, but is it actually necessary?

7. It would be helpful to list the SKOS terms in the table of contents,
perhaps in parentheses.

8. I assume that you have considered naming conventions, for example
skos:prefSymbol versus skos:preferredSymbol.  I don't have a strong
opinion about this, but noticed that abbreviations seem to be
inconsistent.

9. In section "Document Properties", "changes to a concept" should be
"changes to the definition of a concept".  :)

10. In section "Semantic Relationships", I don't know why
"(paradigmatic)" appears.  It does not add anything for me.

11. In section "Concepts in Multiple Schemes":
s/meaning a concept/meaning of a concept/

12. Typo: s/An concept/A concept/

13. Suggeston: Use "http://example.com" instead of
"http://www.example.com", because it's shorter.  :)

Thanks,

David Booth, Ph.D.
HP Software
dbooth@hp.com
Phone: +1 617 629 8881
 
Received on Saturday, 5 November 2005 18:46:56 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:17:19 GMT