W3C home > Mailing lists > Public > public-rif-wg@w3.org > March 2010

Re: RIF XML version of OWL RL rules

From: Stella Mitchell <stellamit@gmail.com>
Date: Mon, 8 Mar 2010 20:05:50 -0500
Message-ID: <d64b0f2c1003081705k478c0e1fyddd94f35647e1ed7@mail.gmail.com>
To: Dave Reynolds <dave.e.reynolds@googlemail.com>
Cc: "Public-Rif-Wg (E-mail)" <public-rif-wg@w3.org>
>>In the
>>EBNF for IRIMETA I think "IRICONST?" needs to be replaced by "Const?".
>
>I don't think so. The intention, I belive, is that rules are identified by
IRIs - you cannot use any sort of CONST. In BLD, section >2.6.3 explains
what an IRICONST is and says the DTB shortcuts are allowed which is why I
believe the angle-bracket notation is >legal.

It would be better if IRICONST were defined in the EBNF so that the EBNF is
self-contained and so that all the symbols are defined in the grammar
language (I think it used to be but got lost in some renaming and
reorganization).

 Also, FLD, both in the mathematical english and in the EBNF, says that the
id part of an annotation is a constant (without restricting it to be rif:iri
constant), and the direct specification of BLD says that the id part is a
rif:iri constant, but the specification of BLD as a restriction of FLD
doesn't mention that restriction.

Stella



On Mon, Mar 8, 2010 at 12:07 PM, Dave Reynolds <
dave.e.reynolds@googlemail.com> wrote:

> On 03/03/2010 12:09, Stella Mitchell wrote:
>
>> Here are the RIF/XML (and associated PS) files [1] for the rules given
>> in Appendix 6 of the OWL RL document [2].
>>
>
> That's great, thanks Stella.
>
>
> I removed all the annotations. Are those in correct syntax?
>>
>
> Ah. Apparently not, sorry about that. (Wonder if they were right at the
> time I first wrote the translation?)
>
> I've changed them to be relative IRIs of the form
>
>    (* <#rule-name> *)
>
> which I think is now right.
>
>
> In the
>> EBNF for IRIMETA I think "IRICONST?" needs to be replaced by "Const?".
>>
>
> I don't think so. The intention, I belive, is that rules are identified by
> IRIs - you cannot use any sort of CONST. In BLD, section 2.6.3 explains what
> an IRICONST is and says the DTB shortcuts are allowed which is why I believe
> the angle-bracket notation is legal.
>
>
> In the SimpleTriple and Datatype rules, I changed all occurences (10 in
>> SimpleTriple, 32 in Datatype) of Foralls without variables, which I
>> think are syntactially incorrect, from (e.g.):
>>        Forall (rdfs:label[rdf:type->owl:AnnotationProperty])
>>   to
>>        rdfs:label[rdf:type->owl:AnnotationProperty]
>>
>
> Thanks. I never could get my head round why we require variable lists at
> all and then don't allow nullary ones. Now fixed in the document.
>
> Dave
>
Received on Tuesday, 9 March 2010 01:06:23 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 9 March 2010 01:06:24 GMT