W3C home > Mailing lists > Public > public-rif-wg@w3.org > July 2008

Re: SWC review part 4 (final part) - (Re: RDF and owl compatibility document ready to be frozen)

From: Axel Polleres <axel.polleres@deri.org>
Date: Thu, 10 Jul 2008 12:07:16 +0100
Message-ID: <4875ED64.6030502@deri.org>
To: Jos de Bruijn <debruijn@inf.unibz.it>
CC: Sandro Hawke <sandro@w3.org>, public-rif-wg@w3.org

ok, all fine with me now! From my side, we can go with SWC.

Jos de Bruijn wrote:
> Axel,
> 
> Thanks for your comments.
> I think the changes prompted by the comments fall in the category of 
> "fixing obvious errors", and so can still be considered for last call. 
> There were one (English) grammatical error, two notational errors 
> (Datatype versus DATATYPE and the repetition in the definition of I* in 
> the proof), and two small technical errors (meta-variables and the 
> reference to the wrong column in the table).
> 
>> Section 8.2.3.2
>>
>> *) "In the following, let T be the set of considered datatypes union 
>> the set of datatypes used in any ontology under consideration."
>>
>> this reads badly, better.. the word "union" is used as in a 
>> mathematical formula in natural language text, better write:
>>
>> "In the following, let T be the union of the set of considered 
>> datatypes with the set of datatypes used in any ontology under 
>> consideration.
> 
> ok
> 
>>
>> *) you use: pred:isDatatype. but pred:isNotDATATYPE ... capitalization!
> 
> ok
> 
>>
>> *) Table embedding OWL DLP, row 10 onwards. You use two explicit 
>> variables ?x and ?y ... this doesn't work recursively, you need 
>> metavariables!
> 
> You are right that in some cases we need meta-variables, but not 
> everywhere.  Namely, in the translation of subclass statements we can 
> simply introduce a new variable, because the statement is translated to 
> a new rule.
> 
>>
>> i.e. change:
>>
>> row 10, column 2: "trO(description1,description2,?x)"
>>
>> to trO(description1,description2, ?<i>x<sub>new</sub></i>)
>>
>> and add a to the condition column the remark:
>>
>> "where <i>x<sub>new</sub> is a "fresh" variable name not used anywhere 
>> else in the translation so far."
> 
> This is not necessary, because the statement is translated to a new rule.
> 
>>
>> analogously, in row 11:
>>
>> instead of "?x" write "?<i>x<i>"
>>
>> and add to the condition column: <i>x<i> is a variable name.
> 
> Yes, here it is necessary.
> 
>>
>> row 12, 13, 14,15, 16:
>>
>> again, you need a metavariable for x.
> 
> In 12 and 13 it is not necessary, but in the other rows it is
> 
> By the way, I realize now that it probably makes sense to move down the 
> row 11 to below what is now a row 13 (it reads more natural) , but this 
> should be done after last call.
> 
>>
>> row 16:
>>
>>   tr<sub>O</sub>(value<sub>i</sub>)
>> should be
>>   tr(value<sub>i</sub>)
>> or no?
> 
> yes
> 
>>
>> row 17:
>>  again meta-variable for ?x .. otherwise, i.e. if you don't use 
>> meta-variables, the other rules don't apply for
>>   tr<sub>O</sub>(description, ?y)
>> you want rows 11-17 also apply recursively to ?y, or no?
> 
>   12 and 13, no, the others, yes
> 
>>
>>
>> *) In the proof of the Normalized Combination Embedding Lemma, this 
>> looks strange to me:
>>
>> "
>> I* = <TV, DTS, D*, D*ind union (union of the value spaces of all 
>> datatypes in the range of D), Dfunc, IC, IV, IF, Iframe', ISF, Isub, 
>> Iisa, I=, Iexternal, Itruth> is such that
>>
>>     * D*ind=Dind union (union of the value spaces of all datatypes in 
>> the range of D) and
>>     * D*=D union D*ind
>> "
>>
>> if "D*ind=Dind union (union of the value spaces of all datatypes in 
>> the range of D)" then why do you define in I*
>> "D*ind union (union of the value spaces of all datatypes in the range 
>> of D)"
>> this is superfluous, it seeems you can just write:
>>
>> "I* = <TV, DTS, D*, D*ind, Dfunc, IC, IV, IF, Iframe', ISF, Isub, 
>> Iisa, I=, Iexternal, Itruth> is such that [...]"
> 
> yes
> 
>>
>> *) Also in the proof further down, you talk about:
>> "(cf. the right column of Table Normalizing OWL DLP)"
>> This table has three, columns... you don't mean the right (condition) 
>> column, but the second (middle) column here, I strongly assume.
> 
> yes
> 
>>
>> Apart from that, all looks quite like an argument with a red line in 
>> the proof to me, although I didn't check back in detail with the OWL 
>> semantics here, I trust you.
> 
> thanks for your confidence :-)
> 
> 
> Jos
> 


-- 
Dr. Axel Polleres, Digital Enterprise Research Institute (DERI)
email: axel.polleres@deri.org  url: http://www.polleres.net/

Everything is possible:
rdfs:subClassOf rdfs:subPropertyOf rdfs:Resource.
rdfs:subClassOf rdfs:subPropertyOf rdfs:subPropertyOf.
rdf:type rdfs:subPropertyOf rdfs:subClassOf.
rdfs:subClassOf rdf:type owl:SymmetricProperty.
Received on Thursday, 10 July 2008 11:08:13 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:33:52 GMT