W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > December 2002

Re: handling rdf:value

From: Graham Klyne <GK@NineByNine.org>
Date: Tue, 10 Dec 2002 10:25:48 +0000
Message-Id: <5.1.0.14.2.20021210101705.038d0b60@127.0.0.1>
To: pat hayes <phayes@ai.uwf.edu>
Cc: w3c-rdfcore-wg@w3.org

At 06:09 PM 12/9/02 -0600, pat hayes wrote:
>>At 10:36 AM 12/9/02 -0600, pat hayes wrote:
>>>Well, yes, but rdf:value *is* contextual, but at least this keeps it in 
>>>bounds. I'd prefer to abandon it, but that's apparently not an option.
>>
>>Er, yes.  In the past few days, I've come across a few instances of its 
>>use in RDF schedule data.  Some of which, I think, doesn't conform with 
>>your proposed "abbreviated form" approach;  e.g.
>
>Well, if I understand these examples, it seems to me that they do, at 
>least plausibly. I gather that the value of the rdf:value triples here is 
>a representation of the 'full' value, and that the others are 'pieces' of 
>it broken out for access via RDF reasoning, is that correct? Because if 
>so, that fits perfectly: in this case presumably it would, or at least 
>could, be appropriate to write that 'full' value as the value of the 
>DTSTART property of the event, no?

OK, if you say so (and I don't mean that facetiously).

The reason I expressed doubt was that the rdf:value value did not match the 
value of some other property, which I had the idea was part of your approach.

As for what the example *means*, I have little more information than you 
;-)  However, your explanation concurs with my understanding.


>>[[
>><VEVENT>
>><!--- snipped -->
>><DTSTART>
>>
>><DATE-TIME>
>><TZID rdf:resource="#US-Eastern"/>
>><rdf:value>20010226T090000</rdf:value>
>><util:hour>09</util:hour>
>><util:minute>00</util:minute>
>></DATE-TIME>
>></DTSTART>
>>
>><DTEND>
>><DATE-TIME>
>><TZID rdf:resource="#US-Eastern"/>
>><rdf:value>20010227T173000</rdf:value>
>><util:hour>17</util:hour>
>><util:minute>30</util:minute>
>></DATE-TIME>
>></DTEND>
>></VEVENT>
>>]]
>>-- http://www.ilrt.bristol.ac.uk/discovery/2001/06/content/rdf_meeting.rdf
>>
>>This is just an example picked at random.  I've noticed this pattern a 
>>couple of times in iCalendar/RDF data.
>
>Look, there is a more general issue here. Our charter asks us to make the 
>RDF M&S clear. That thing that we were supposed to clarify has spawned 
>this monstrosity BY BEING UNCLEAR about what rdf:value is supposed to mean 
>or to be appropriately used for. The result, that existing code use it for 
>all kinds of not-mutually-compatible things, is the PROBLEM that we are 
>supposed to be solving, seems to me, not the state of affairs that we are 
>supposed to preserve by avoiding anything that might violate some use 
>case. If the existing use cases, taken as a whole, are confused, then one 
>or more of them have GOT to change, or else we just declare that confusion 
>reigns and we are not going to do anything about it. Which, to return to 
>my first point, seems to me to be a clear violation of our charter. If 
>confusion were supposed to reign, there would have been no need to form 
>this WG in the first place.

Yes, no argument from me...

>So I think that we should say something that makes an intended use case 
>for rdf:value reasonably clear, and then say that this is what rdf:value 
>is supposed to mean. We should try to capture as many use cases as we can, 
>of course, but if we can't get them all, we should not just give up. Im 
>not sure what the above pattern is supposed to mean, to tell you the 
>truth, but if it can't be interpreted in the way I suggest, then 
>iCalendar/RDF will have to invent a new property: tough.

... if your explanation covers this case, then I think there's no more to 
say, and I'm pleased that the data I found is not an impediment to your 
proposal to describe the intended meaning of rdf:value.

#g


-------------------
Graham Klyne
<GK@NineByNine.org>
Received on Tuesday, 10 December 2002 05:33:57 EST

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:54:50 EDT