[OEP] Re: Typo in XML Representation for Approach 3 in CAV

(Chris brought this to my attention in the context of 'can we edit
this in place?'  I'm not sure of the answer yet to Chris' question,
as it depends on how much the WG wants to change.)

At 04:32 PM 6/22/2005 -0400, Steven Wartik wrote [1]:

>Figure 3 has a class named "Book". However, the RDF/XML for approach 3 has a class with ID "BookAboutAnimals".

   [1] http://lists.w3.org/Archives/Public/public-swbp-wg/2005Jun/0069.html

There are several discrepancies between the figures, the N3 text
in the body of the Classes As Values Note [2], and the associated
N3 and RDF/XML files.  Steve (mis) identified one issue but there
are others.

   [2] http://www.w3.org/TR/2005/NOTE-swbp-classes-as-values-20050405/

I don't claim the following is a complete list of issues with the
.n3 and .owl code files:

In Approach 1 the N3 in the body of the document includes two
rdfs:seeAlso statements that are not shown in Figure 1.  (The N3
also includes :bookTitle properties that are implicit in the figure.)
These seeAlso statements do not appear to add much to the
example.  I also suspect that better practice for the intended
semantics would be to use dc:identifier rather than rdfs:seeAlso.
One of the two books has an owl:sameAs statement giving
the same object URI; this almost certainly should be dc:identifier
rather than owl:sameAs.

books2.n3, books2.owl, books3.n3, and books3.owl also contain these
unnecessary rdfs:seeAlso statements.

books1.n3 has

  @prefix default:  <http://protege.stanford.edu/swbp/books1.owl#> .

whereas books1.owl has

   xml:base="http://protege.stanford.edu/swbp/books1.owl"

(note the missing '#' -- this is a bug.  Probably best to avoid
xml:base for this purpose and use an xmlns declaration explicitly.)

books3.n3 fails to force an effective base URI of
http://protege.stanford.edu/swbp/books1.owl#
so it does not generate precisely the same triples as books3.owl

The real typo in books3.n3 and books3.owl (also books2.n3 and
books2.owl) is that BookAboutAnimals is not explicitly declared to be rdfs:subClassOf :Book to follow with Figures 2 and 3.

I think the approach 3 range restriction on dc:subject would be
more clear if the class AnimalSubject were defined, with that
being the parentSubject of LionSubject.  That would then be
the full analog of the approach 1 and 2 examples.

So I'm not sure how to respond to the comment in [1].  It's a bit
off the mark but it did point out some true flaws in the N3 and
RDF/XML files.

Received on Wednesday, 21 September 2005 20:04:09 UTC