Re: Well Behaved RDF - Taming Blank Nodes, etc.

Thanks all, especially Pat.
That all makes some sense to me - I was despairing that I understood, and that the discussion might reach any consensus.

So no-one is suggesting outlawing blank nodes.
But there is something of a feeling that having a URI is a useful thing, and if it isn't too much trouble, please can you use a URI.

After all, the work to publish RDF only gets done once, but (hopefully!) the work to consume it gets done a lot.


I hesitate to go back to the "complete nonsense", but I think David's example actually also illustrates the issue I raised with labels and strings.
> A string literal should only ever appear as the object of rdfs:label (or something that could naturally be made a sub-property of rdfs:label, such as foaf:name).


Consider what David might have said:

:some-photo :depicts "Horse standing next to Ann Romney 2012-07-25" .

It seems much better to say
:some-photo :depicts :romney-horse-2012-07-25 .
:romney-horse-2012-07-25 rdfs:label "Horse standing next to Ann Romney 2012-07-25" .

For exactly the same arguments about blank nodes v. URIs.
But people do have a tendency to ignore that there is a horse resource. Making them ask the question of whether there should be a URI for the resource, as a rule of thumb, is in my view a useful thing to do. You can always decide not to.

I actually think it is bad modelling to use one triple, but as I said before, I see it all the time.
It is using a resource, which is a string, as a proxy for something else.
A major promise of the Semantic Web (and especially Linked Data) was that we would get away from using strings as identifiers for things, with their inherent ambiguity &c..

I am sure some people will still consider this "complete nonsense", but I find it really useful in terms of generating my RDF, as well as advising other people how to make their RDF easily consumable by me. I don't always introduce a resource, but it means I am less likely to miss one out accidentally.

There you go.
Best
Hugh

On 16 Dec 2012, at 04:05, David Booth <david@dbooth.org> wrote:

> Hi Pat,
> 
> On Fri, 2012-12-14 at 23:52 -0800, Pat Hayes wrote:
>>>>> On 12/13/2012 2:00 AM, Pat Hayes wrote:
>>>>>> Another example: a picture of some celebrity standing next to a
>> horse. I have a URI for the celebrity, but I don't have and don't need
>> one for the horse [ . . . ]
> 
>> I don't need, and shouldn't have to invent, a URI for something that
>> does not need to be identified. Particularly if in fact I cannot
>> identify the thing in quesiton. Inventing a URI to "identify" an
>> anonymous horse or an unknown room in a building is simply a mis-use
>> of the very idea of a URI as being a universal identifier. 
> 
> [The following comment is moot with respect to Well Behaved RDF, because
> as Lee showed, you could use Well Behaved RDF and still have an implicit
> blank node for the horse, but . . . ]
> 
> I can see that you may find it inconvenient to mint a URI for the horse,
> but I vehemently disagree with the idea that it would be a mis-use of a
> URI.  As the AWWW states:
> http://www.w3.org/TR/webarch/#uri-benefits 
> "A resource should have an associated URI if another party might
> reasonably want to . . . make or refute assertions about it . . . ."
> The fact that *you* chose to make an RDF statement about that horse is
> pretty strong evidence that someone else may eventually want to make
> statements about it also.  One doesn't have to know a lot about a
> resource to mint a URI for it.  Something like this would suffice:
> :some-photo :depicts :ann-romney ,
>     :romney-horse-2012-07-25 .
> :romney-horse-2012-07-25 a :Horse ;
>     rdf:label "Horse standing next to Ann Romney 2012-07-25" .
>

Received on Sunday, 16 December 2012 14:59:59 UTC