Re: ACTION-420 Review of SW-compatibility

Axel,

Thanks again for the review.  Replies on selected comments that we did 
not discuss at the face-to-face in line.

> 17)
> 
> "IR union IP"
> 
> better:
> 
> "IR ∪ IP"

I disagree.  Most browsers are bad at rendering mathematical symbols; I 
believe some don't render them at all.

> 
> 18)
> "Rdf-, rdfs-, "
> 
> -->
> "RDF-, RDFS-, "
> check capitalization in the whole document

  I am following the conventions in the RDF semantics specification.

> 
> 19)
> In te definition of common interpretations, isn't condition 2 superfluous? (by condition 4.)

did you mean "by condition 5"? The domain of IEXT is IP.  If it were not 
for condition 2, IEXT might be ill-defined.

> 
> 21)
> "i.e., a rdfs:subClassOf b is true if a ## b is true."
> 
> Why is "if" enough here, and not "iff"? not sure.

Well, this condition is phrased as per the working group resolution [1]. 
  The reason the text is phrased as it is, is that ## is meant to be 
irreflexive, whereas rdfs:subClassOf is reflexive.


[1] RESOLVED: Close Issue-43 by including in BLD subclass formulae of 
the form a ## b. In the RDF compatibility document, ## and 
rdfs:subClassOf will be connected appropriately, i.e. whenever a ## b 
holds, a rdfs:subClassOf b is required to hold. (January 8)

> 22)
> What exactly is meant by
> "which are used as properties in the RIF domain. "
> ?

With "properties in the RIF domain" I mean the objects in the domain of 
the function I_frame, intuitively the things occurring in the p-position 
of frames s[p -> o].
Do you have a suggestion for a better phrasing?

> 
> 23)
> I don't really understand here:
> 
> "Note that no correspondences are defined for the mapping of names in RDF which are not symbols of RIF, e.g., ill-typed literals and RDF URI references which are not absolute IRIs."
> 
> That means that what is written in the example before breaks:
> 
> "
>>From this combination we can derive the RIF condition formulas
> 
> Exists ?z ( ?z[rdf:type -> "nameBearer"^^rif:iri] )
> Exists ?z ( "http://a"^^rif:iri["http://p"^^rif:iri -> ?z] )
> 
> as well as the RDF triples
> 
> _:y rdf:type nameBearer .
> <http://a> <http://p> "a"^^xsd:integer . 
> 
> However, "http://a"^^rif:iri["http://p"^^rif:iri -> "a"^^xsd:integer] cannot be derived,
> "
> 
> By what means is the triple
> "<http://a> <http://p> "a"^^xsd:integer . "
> derived in this example, if not from the RIF model theory??? 

The point I was trying to get across is that there is no syntactical 
representation in RIF for ill-typed literals.
Do you have a suggestion how I could make this clearer in the text?


> 26)
> "We say that I DL satisfies a rule Q "
> 
> shouldn't that rather be:
> 
> "We say that I DL satisfies a DL rule Q "
> ??? (however, we only define DL rulest, not DL rule before!)

I believe it is not necessary to define DL rules (this saves us from 
having to introduce another concept).  DL-satisfaction is well defined 
on all rules. I did change the definition to define DL-satisfaction only 
for DL-rule sets.


Best, Jos

-- 
                          debruijn@inf.unibz.it

Jos de Bruijn,        http://www.debruijn.net/
----------------------------------------------
One man that has a mind and knows it can
always beat ten men who haven't and don't.
   -- George Bernard Shaw