Re: URIs within URIs

Hi Luca,

> I'm wondering however if *form URIs* could themselves be resources
> rather than datasets.

Sure, why not. They're just URIs.

> For example imagine the example from earlier:
> 
>    <http://bar.com/?subject=http%3A%2F%2Ffoo.com%2Falice> a void:Dataset .
>    <http://foo.com/alice> <#some> <#data> .
> 
> Instead of being a void:Dataset, the form URI could simply be a new
> URI someone defines with an owl:sameAs relation (or other outbound
> relationships) to the alice URI, such as:
> 
>    <http://bar.com/?uri=http%3A%2F%2Ffoo.com%2Falice> owl:sameAs
> <http://foo.com/alice>;
>                <#some> <#data> .

We have to be careful with use/mentions though.
Is bar.com… really Alice, or a document about Alice?
And is foo.com/alice a document, or Alice?

Only if both are Alice, the sameAs makes sense.
If both are documents, they're probably not the same;
if either is a person, then they're certainly not the same.

[Note to readers: please don't restart httpRange-14 here.]

> I guess the differentiation between the two is a simple semantic
> formality. Can I link to triple pattern fragments datasets as if they
> were resources? Such as: <http://other.com/alice> rdfs:seeAlso
> <http://bar.com/?subject=http%3A%2F%2Ffoo.com%2Falice> . ?

Absolutely. Note that similar constructs are already present in fragment representations:

http://data.mmlab.be/mmlab?subject=http%3A%2F%2Fdata.mmlab.be%2Fpeople%2FRuben%2BVerborgh

contains

<http://data.mmlab.be/people/Ruben+Verborgh> rdfs:seeAlso <http://data.mmlab.be/mmlab?object=http%3A%2F%2Fdata.mmlab.be%2Fpeople%2FRuben%2BVerborgh>.

(And note the difference between the person and the document.)

Best,

Ruben

Received on Monday, 25 August 2014 11:27:00 UTC