Search Mailing List Archives
[p4-feedback] Problems with datatypes in class expressions.
nick.drummond at cs.manchester.ac.uk
Thu Apr 30 06:19:16 PDT 2009
2009/4/28 S Essaid <sielists at gmail.com>
> Hi Nick,
> I noticed few problems with the class expression editor's handling of
> datatype declarations.
> - (dataproperty some rdfs:Literal) will be shown as (dataproperty some
> Literal) when the ontology is reopened. If any part of this description is
> edited, the Literal (now without the rdfs:) will be assumed to be
> xsd:Literal. This is not a valid datatype and it is a change from what it
> originally was. The only way to avoid this is to reenter all the prefixes to
> rdfs:Literal and rdf:XMLLiteral. The editor converts everything to the xsd
OK, this only happens when a qname renderer is used. I've fixed it.
> - (dataproperty max 1 rdfs:Literal), the datatype is not saved in the
> ontology (when looking at the RDF/XML) but this is consistent with
> rdfs:Literal being the default type. The problem is that when the ontology
> is reopened, this default datatype is shown as "Literal" in the expression
> editor and it will be saved as xsd:Literal if the description is edited and
> the "Literal" is not prefixed or removed.
Above fixes this
> - The expression editor does not provide the datatype completion list when
> a datatype qualified cardinality restriction is being entered.
OK, I fixed this for object properties a while back and obviously missed
this. Added to the list.
> I found a previous post where similar issues were noted.
Yeah I fixed it once for the general case - was just an error in the object
> I really need this to work because I am writing many descriptions with
> rdfs:Literal and it is becoming difficult to reedit these descriptions.
> Also, this problem is affecting Pellet by causing it to mark many classes as
> unsatisfiable when the rdfs:Literal is accidentally converted to
> xsd:Literal. It took me several hours yesterday to figure out what was going
> on because I am new to predicting inferences and I thought that I was doing
> something wrong but the culprit turned out to be the incorrect xsd:Literal
> declarations that are only noticeable when examining the source.
Do you need the qname renderer or could you get away without temporarilly
until the next build?
> Thank you,
> p4-feedback mailing list
> p4-feedback at lists.stanford.edu
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the p4-feedback