Re: Review of the QCR draft

Jacco

I had not thought of adding a formal semantics section because QCRs  
are standard constructs in DLs with a standard model theoretic  
semantics.  Rather than include the full mathematical notation for  
the model theoretic semantics, which would be very much at variance  
with the style of the other notes, I'd rather add a pointer to the DL  
Handbook and/or other references.

I agree about the RDF triple stuff.  I had included that following  
Guus' original note.  I think you are right that it stands out.  At  
the very least the abstract syntax should be first with a note in the  
discussion showing what it means in terms of RDF triples.

Would these two changes fit your suggestions?

Alan

On 5 Nov 2005, at 19:48, Jacco van Ossenbruggen wrote:

>
> Exec summary:
>
> I think the draft could be published as a note if comments below  
> are addressed, only I'm worried about the lack of a formal  
> semantics of the owl extentions proposed.
>
> Jacco
>
> Review of the QCR draft
> Qualified cardinality restrictions (QCRs):  Constraining the number  
> of values of a particular type for a property
> W3C Editor's Draft 02 November 2005
> http://lists.w3.org/Archives/Public/public-swbp-wg/2005Nov/att-0022/ 
> QCRs-revised-01.html
>
>
> General remarks
> ===============
> Draft is very clear until pattern 3
> - "The syntax looks like this [3]."  This left me confused.  Does  
> the reader really need to go to [3] to look up what the syntax  
> looks like?
>  Please remove this sentence if not important or give examples of  
> the proposed syntax inline.
> - The mechanism in the ordered list items 1,2, 3 is explained on  
> the RDF triple level, but the example syntax is given in the  
> abstract syntax,
>  which makes this unnecessary complex.  I suggest to first explain  
> everthing in terms of the abstract syntax, and than discuss what this
>  means on the RDF triple level, including all triples involved in  
> the example in turtle or n3 notation, higlighting the "extra" ones.
> - "where has_level and has_trend are functional properties but  
> has_feature is not functional.
>  Why? Is this relevant for the QCR?  Should these properties be  
> declared functional in the example?
>
> Discussion section:
> - "The intended semantics fully capture the intent of the use cases  
> given."
>  Sure.  But what about the acual semantics?  Is there a formal  
> semantics for this extention? Please refer to this!
> - "Simple cardinalty restrictions can be taken as a special case of  
> Qualified Cardinality Restrictions where qualifying argument is  
> valuesFrom(owl:Thing)."
>  So what?  Please explain, give an example, state if this is good  
> or bad, or only interesting from an academic point of view.  Or  
> remove?
>
> Low level stuff
> ===============
> HTML code does not validate and has lots of   onweird places  
> that make the formatting look odd.
>
> <title>:
> - replace title (Cut & paste error from the part/whole draft)
>
> Copyright:
> - 2004 or 2005?
>
> Status:
> "This document will be a part of a larger document that will  
> provide an introduction and overview of all ontology design  
> patterns produced by the Semantic Web Best Practices and Deployment  
> Working Group."
> - Be more specific and link to this document if it exists.  I  
> suggest to remove this remark if doc does not exist.
>
> Basics:
> - s/valujes/values
>
> Use cases:
> Rector [1]
> - Make reference into a hyperlink, also for all other [1], [2], [3]  
> refs in document
>
> use case 2
> - s/opne/one
>
> use case 6
> - s/undertainty/uncertainty
> - s/lline/line
>
> Representation Pattern 1: Use owl:someValuesFrom:
> -s/generalisations of 3/generalisations of use case 3
>
> Representation Pattern 2: 'Work around" using subPropertyOf
> - make quotes in 'Work around" heading consistent (remove switching  
> between " and ' in entire document)
> - s/hve/have
> -
>
> Representation Pattern 3: Use a non-endorsed OWL extension
> - first line: s/QCR/QCRs
> - s/suggestd/suggested
> - s/incoporatd/incorporated
> - "OWL users who require QCRs for use cases not well served by the  
> work around in Representation 2 may want to use this extension"
>  Rephrase or add commas?
> - "At least one widely used tool already supports QCRs as do many  
> of the widely used classifiers."
>  Add comma after QCRs
> - "To this a qualified cardinality constraint need only add a third  
> triple"
>  Syntax looks odd for a non native speaker like me (only needs to  
> add a third triple?)
>  Similar phrasings in the rest of this para.
> - "For example the exampole"
>  Fix typo and rephrase using "example" only once.
> - "but each Feature could"
>  lower caps F
>
> Discussion:
> - replace [Note 4] by link to Note [1]
>

-----------------------
Alan Rector
Professor of Medical Informatics
Department of Computer Science
University of Manchester
Manchester M13 9PL, UK
TEL +44 (0) 161 275 6188/6149
FAX +44 (0) 161 275 6204
www.cs.man.ac.uk/mig
www.clinical-esciences.org
www.co-ode.org

Received on Sunday, 6 November 2005 11:23:08 UTC