W3C home > Mailing lists > Public > public-rdf-dawg@w3.org > January to March 2010

Re: ISSUE-54: Do we need (descriptions of) property functions in SD? Is this in scope for us?

From: Steve Harris <steve.harris@garlik.com>
Date: Wed, 17 Feb 2010 09:54:24 +0000
Message-Id: <837AECC9-1F80-4930-87F5-E2A4559B2033@garlik.com>
To: SPARQL Working Group WG <public-rdf-dawg@w3.org>
On 17 Feb 2010, at 05:33, Paul Gearon wrote:
> On Tue, Feb 16, 2010 at 11:15 PM, Gregory Williams
> <greg@evilfunhouse.com> wrote:
>> On Feb 16, 2010, at 12:28 PM, Steve Harris wrote:
>>
>>> On 16 Feb 2010, at 17:23, Lee Feigenbaum wrote:
>>>
>>>> I'd be surprised if anyone disagrees that this is useful.
>>>>
>>>> However, I have no idea how we would specify it in the service  
>>>> description document. How would we define a property/class that  
>>>> describes something that is not itself defined anywhere? The only  
>>>> way I see to do it is to define what a property function is, and  
>>>> that's beyond our scope.
>>>
>>> Right, this is my feeling too.
>>
>> Can I take this to mean that my suggested wording isn't acceptable?  
>> I'm not sure we need to define exactly what happens when a property  
>> function does its thing so long as we indicate that it's up to the  
>> implementation, but others may think differently.
>
> I was happy with it, but apparently something is missing.
>
> I think it's clear that we can't (or shouldn't try to) define what a
> property function is. I just want to see the particular predicates
> listed. That's trivial to do, and it's all that I want.

Perhaps I'm missing something, but looking at it from a triply point  
of view, you have something like:

:propertyFunction rdfs:range :PropertyFunction .
<endpoint> :propertyFunction <func> .

But, you can't define what a :PropertyFunction is, as it's not a  
standardised thing, so two people reusing the same <func> URI is at  
best unwise, and in the worse case dangerously misleading, and the  
implied typing is meaningless.

So, I don't see what this achieves, or how we could justify including  
it in a rec.

However, we're using RDF for our descriptions, so nothing stops a  
group of implementers agreeing on what a "Property Function" is  
outside this group, and adding some properties and classes to an  
appropriate schema.

- Steve

-- 
Steve Harris, Garlik Limited
2 Sheen Road, Richmond, TW9 1AE, UK
+44 20 8973 2465  http://www.garlik.com/
Registered in England and Wales 535 7233 VAT # 849 0517 11
Registered office: Thames House, Portsmouth Road, Esher, Surrey, KT10  
9AD
Received on Wednesday, 17 February 2010 09:54:54 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:15:41 GMT