Re: changes on rdf:text on the remaining Ed notes

Ian Horrocks wrote:
> Alex,
^^^^^^ Axel ;-)

> I wonder if it isn't an unnecessary burden on OWL and RIF to be trying 
> to define stuff which is not relevant to either OWL or (if I understand 
> you correctly) RIF. From what you say, these functions are for possible 
> use by future XPath/XQuery implementations that support rdf:text. But in 
> this case, couldn't/shouldn't we remove them from this spec and leave it 
> up to a/the XPath/XQuery WG to come up with a suitable design?
> 
> Ian

The rationale is: RIF built-ins are to a large extent just adaptions of 
the respective XPath/XQuery built-ins. That way, an XPath/Xquery library 
should be re-usable for RIF implementations. In XPath/Xquery there is 
functions and operators defined for the primitive datatypes. What we do 
is a new primitive datatype, and as such, I suggest to have a set of 
basic functions and ops defined in that same way, re-sable for RIF, as 
well as for extensiond of XML query languages. In my opinion, as 
definers of the datatype, this is well inside our scope.
If we do it in a RIF-only way, I personally think it is too narrow.

Also, it seems we are there already, the only thing missing being a 
namespace. Something like

rdftfn: http://www.w3.org/2009/rdftext-functions

would be fine, it seems.





> 
> 
> 
> On 5 Apr 2009, at 22:33, Axel Polleres wrote:
> 
>> Sandro Hawke wrote:
>>>> Alan Ruttenberg wrote:
>>>>> Hi Axel,
>>>>>
>>>>> What are the plans for the namespace for these functions? I'm
>>>>> concerned any resolution of that will delay publication of the
>>>>> document, and OWL plans to go to CR in the not to distant future and
>>>>> has a dependency on it I'd like to minimize risk on that.
>>>>>
>>>>> Have you considered moving these functions out of the rdf:text
>>>>> specification and into RIF instead?
>>>> I am reluctant on this one, to be honest: These functions are not 
>>>> RIF-functions but XQuery/XPath style functions. So, I don't see why 
>>>> and in which RIF spec they should go.
>>>>
>>>> I agree that "hijacking" the fn namespace is problematic. Can we 
>>>> give it our own namespace? e.g. analogously to
>>>>
>>>> fn: http://www.w3.org/2005/xpath-functions
>>>>
>>>> we could use
>>>>
>>>> rdftfn: http://www.w3.org/2009/rdftext-functions
>>>>
>>>> That would be the simplest solution, IMO.
>>> How awkward would to be to leave this undefined?  These URLs are never
>>> used in RIF, and before they could be used in XPath, they'd have to be
>>> adopted by the appropriate WGs, which could put them into the 2005
>>> namespace, I think.  (Or maybe not -- I don't know the XPath
>>> extensibility story.  Do any of us know how these URLs are supposed to
>>> be used?)
>>
>> I guess by future XPath/XQuery implementations that support rdf:text?
>> Not sure, but I'd definitly find it awkward to leave it open, i.e. 
>> you'd suggest to have a standard document which is inherently 
>> incomplete and says, that future standards will complete it (by 
>> defining the resp. namespace)?
>>
>> Axel
>>
>>>     -- Sandro
>>
>>
>> -- 
>> Dr. Axel Polleres
>> Digital Enterprise Research Institute, National University of Ireland, 
>> Galway
>> email: axel.polleres@deri.org  url: http://www.polleres.net/
>>
> 


-- 
Dr. Axel Polleres
Digital Enterprise Research Institute, National University of Ireland, 
Galway
email: axel.polleres@deri.org  url: http://www.polleres.net/

Received on Monday, 6 April 2009 09:10:41 UTC