Re: Editorial changes to InternationalizedStringSpec

Hi all,

short reporting from the RIF F2F:

We didn't discuss the issue of including constraint facets in rdf:text 
in detail, but - at least with sandro - we came to the following 
conclusion with which I think we can go ahead, if you agree.

Basically, it makes sense to include definitions and constraint facets, 
as well as specific functions and operators with the datatype definition
in our document [2].

So

1) rdf:text could include constraint facets,in the "XML schema style" to
make the facets reusable also for XML as well as for OWL... that was
your intention, right? After rethinking/discussing it, I agree now that
this makes sense, we can, in RIF then just emulate them.

2) rdf:text should then also include some functions and Operators in the
XQuery/XPath functions and Operators [1] style, reusable for us in RIF.

Specifically:

* fn:lang-from-text  (extract the lang-tag as a string)
* fn:string-from-text (extract the string-part)
* op:text-equal     (equality in the sense that both the string part and
the lang-tag are equal)
* fn:lang-matches (lang-tag matches in the spirit of 
http://www.ietf.org/rfc/rfc4647.txt)
* fn:text-from-string-language( s as string, l as xs:language )

...for the latter maybe better to use

* fn:text-from-string-language( s as string, l as xs:string )

since the xs:language datatype refers to RFC3066 and not bcp47? just to 
be on the safe side?


Would those be ok for you?
Other functions such as fn:string-length could be added/adapted, but
seem to be covered by first applying fn:string-from-text and then using
those defined for strings.

* Maybe also some remarks would be in order whether and how
rdf:text fits into the "casting" defined in Section 17.1 of XPath/Xquery
functions and operators [1]. We should decide:

(i) Is text castable to string? I think the easiest solution here would
be to say that only those texts with an empty lang-tag are castable to
xs:strings, all other options seemed to raise concerns.

(ii) an rdf:text cast function which allows to convert xs:strings
and its subtypes to rdf:texts as already mentioned in Section 3.1 of our
document [2]

I think that's all for the moment. If we can agree on that, I think we 
are very close to finishing this!

comments welcome,
Axel

1. http://www.w3.org/TR/xpath-functions/
2. http://www.w3.org/2007/OWL/wiki/InternationalizedStringSpec

Axel Polleres wrote:
> Ivan Herman wrote:
>>
>> Axel Polleres wrote:
>>>> Axel, my reading of the RIF document may be very wrong, bear with me...
>>>> But if I understand it well, RIF's built-ins[7] rely on the definition
>>>> of XPath functions[8], which, indirectly, rely on the facet for a
>>>> specific (xsd) datatype I believe (although I have not found explicit
>>>> reference there:-(.
>>> We only reference specific primitive datatypes at the moment and do not
>>> allow or define facets at the moment. that is, in BLD we do not yet
>>> support arbitrary XSD datatypes, but only a finite, specific set of
>>> these. Actually, each dialect in RIF is associated with the datatypes it
>>> allows. That is, if you wan to support more datatypes, you need to
>>> define a different dialect.
>>>
>>
>> I am sorry, Axel, I was not clear, my bad. When I wrote 'have not found
>> explicit reference there', what I meant was not you or the RIF document,
>> but the XPAth document that does define functions based on XSD, but does
>> not explicitly refer to the constraining facets of XSD when they define,
>> say, less-than. But I would believe that this is what they refer to.
>>
>> But it is true, isn't it, that RIF refers to the XPAth functions for
>> (part of its) built-ins, just like SPARQL does?
> 
> Yes and no, because we have to cast them into our model-theoretic 
> semantics (which e.g. means that we do not prescribe/constrain the 
> error-behavior, but only define the functions' and predicates' 
> interpretations on their "intended domains").
> 
> Axel
> 


-- 
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 Saturday, 27 September 2008 20:49:43 UTC