Re: Review of Requirements Document

Some comments are interspersed below:

Jim Hendler wrote:
> 
> Review of WD version of requirements document
> 
> 1. I don't know if LC documents contain a "change from last version"
> section - if appropriate, be sure to update (bookkeeping comment only)
> 
> 2. As discussed, some of our requirements were not met and should be
> demoted - I  believe Jeff is on top of this.  However, the following
> is how we specified R12, Unique names assumption:
> 
> >R12. Local unique names assumptions
> >
> >In general, the language will not make a unique names assumption.
> >That is, distinct identifiers are not assumed to refer to different
> >objects (see the previous requirement). However, there are many
> >applications where the unique names assumption would be useful.
> >Users should have the option of specifying that all of the names in
> >a particular namespace or document refer to distinct objects.
> 
> I believe we have actually met this if we accept the allDifferent (or
> allDistinct) construct.  I realize this means listing the names in
> the namespace to specify they are unique, but in fact the above
> doesn't state that one shouldn't have to do this.  I consider it like
> where RSS requires the listing of channels, even if they're defined
> in a document, because you both have to have the definitions AND the
> statement that they are to be used.  I thus think it could be argued
> that we should leave R12 as a requirement (I do not feel strongly on
> this)

I think this would go against the spirit of the requirement. I would
rather reword it to explicity rule out the "list all of the names in an
allDistinct construct" option and demote it to an objective, so that it
serves as a place holder for future work by the next OWL WG.
 
> 3. R13 is that we have a way that statements can be "tagged" (quotes
> in the document) with additional information, and stating RDF
> reification may be a way to achieve it.  I believe we have achieved
> this objective only in allowing RDF tagging to be added to our
> graphs, and Jeremy has recently raised some points about whether the
> RDF Graphs that would be produced would be in the OWL syntax for Lite
> or DL documents.   If this is true, we may not have fully met our
> tagging requirement

Do the graph problems occur with OWL Full? If not, then haven't we met
the requirement in OWL Full? I thought it was meant to be a true
superset of RDF?

> 4. R19, R20 - I believe these are met through RDF, but someone needs
> to check. (R18 is clearly met by rdf:label)

I agree (both that I believe they are met and that someone else needs to
confirm).

> 5. We actually have reached a couple of our objectives - perhaps if
> we can demote requirements we could consider promoting objectives?
> If so, I would argue we might move the following to requirements:
> 
> >O1. Layering of language features
> >
> >The language may support different levels of complexity for defining
> >ontologies. Applications can conform to a particular layer without
> >supporting the entire language. A guideline for identifying layers
> >may be based on functionality found in different types of database
> >and knowledge base systems.
> 
> we clearly achieved this.
> 
> >O6. Effective decision procedure
> >
> >     The language should be decidable.
> 
> We have defined a decidabe and efficient subset of our language (OWL
> Lite), and a decidable subset (OWL DL).  I believe this could allow
> us to move 06 to a requirement that read more like follows:
> 
> Rxx. Effective Decision procedure
> 
>    Many applications may require specific application guarantees for
> reasoners including decidability with an effective decision procedure
> or just decidability.  Although general ontology reasoning is
> undecidable, subsets can be defined which have these types of
> specific guarantees, and the Web ontology language should identify
> and support such subsets.
> 
> [note, Ian or someone might want to fix my wording above if I didn't
> get the details right]

I would be willing to make these changes if the WG thinks its a good
idea.

> --
> Professor James Hendler                           hendler@cs.umd.edu
> Director, Semantic Web and Agent Technologies     301-405-2696
> Maryland Information and Network Dynamics Lab.    301-405-6707 (Fax)
> Univ of Maryland, College Park, MD 20742          240-731-3822 (Cell)
> http://www.cs.umd.edu/users/hendler

Received on Thursday, 2 January 2003 11:32:58 UTC