Re: How do you deprecate URIs? Re: OWL-DL and linked data

On Jul 16, 2008, at 6:49 PM, Booth, David (HP Software - Boston) wrote:

> Hi Bijan,
>
>> From: Bijan Parsia
>>
>> On 15 Jul 2008, at 18:15, Booth, David (HP Software - Boston) wrote:
>>
>>> It is the *URI* that you wish to deprecate -- not the resource
>>> denoted by that URI, so the deprecation statement *must*
>>
>> Not really.
>>
>>> be written something like this:
>>>
>>>   "http://lingvoj.org/iso-639/languages#da"^^xsd:anyURI
>>>       owl:deprecatesURI
>>>
>> "http://downlode.org/rdf/iso-639/languages#da"^^xsd:anyURI  .
>>>
>>> using "" rather than <> to indicate that you are talking about the
>>> URI itself, rather than the resource denoted by that URI.
>>
>> You could easily figure it out from context.
>>
>>> The Semantic Web community does not yet seem to be accustomed to
>>> taking about URIs themselves,
>>
>> People make use/mention slides all the time. In all sorts of context.
>> Sometimes it matters. Sometimes it doesn't. I don't know which it is
>> in this case.
>
> Certainly if you step outside of RDF

What? I'm just pointing out that syntactic explicit disambiguation  
isn't the only way to handle use/mention and that furthermore, people  
are good (in some context) at dealing with it. It's not clear that  
there's a problem with this particular use/mention conflation such  
that we need to train ourselves out of it.

> there are many ways the distinction can be made.  The discussion (I  
> thought) was about how to make a deprecation statement *in* RDF.

Sure.

> RDF has no notion of context.

Of course it does, otherwise your above trick wouldn't work. The  
string inside the quotation markes is a literal and the one in <> is  
an identifier. (This is a serious point, by the by.) Furthermore, you  
can distinguish things on the triple level if you are willing to have  
a difference in some part of the vocabulary.

In other words: C rdf:type C. has at least two contexts (the subject  
of an rdf:type and the object of an rdf:type). And there's no  
ambiguity. That's my point above. You could just say C rdf:type  
DeprectatedURI was sugar for the literal.

I'm not saying it's the most pleasing, but it may be more usable.  
People have been yelling about such things for aeons (witness the  
famous x dc:creator "bijan" "BIJAN ISN'T A STRING" debates which,  
imho, seriously hurt dublin core and thus RDF adoption at a critical  
phase; or heck, hash vs. slash).

Look, I love being the pedant as much as anyone (if it weren't  
obvious; it's a pretty deadly and telling insult in philosopher land  
to accuse someone of use/mention confusion), but it's just not clear  
to me that this bit of pedantry is worth it.

Seems like it would be better to give a big chunk of functionality  
instead.

Cheers,
Bijan.

Received on Thursday, 17 July 2008 20:39:00 UTC