parsing lists of CURIES

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
3 messages Options
Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

parsing lists of CURIES

Dan Connolly
how many CURIEs in this attribute?

 typeof="foaf:Person ***foaf:Agent 123412:abc owl:Thingfoaf:Agent"

I don't see anything in the spec about how to split an
attribute value into a list of curies.

Do I split by whitespace and then throw out the non-curies?
(such as ***foaf:Agent, leaving just 1 CURIE, foaf:Person)

Or do I find all occurrences of CURIE syntax?
i.e. 4 CURIES: foaf:Person, foaf:Agent, owl:Thingfoaf
and Thingfoaf:Agent?

or all non-overlapping occurrences? i.e. 3

What if foaf: is bound but owl: isn't?

Are there relevant test cases? I hope we come up with some.

--
Dan Connolly, W3C http://www.w3.org/People/Connolly/
gpg D3C2 887B 0F92 6005 C541  0875 0F91 96DE 6E52 C29E


Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: parsing lists of CURIES

Mark Birbeck-4
Hi Dan,

> how many CURIEs in this attribute?
>
>  typeof="foaf:Person ***foaf:Agent 123412:abc owl:Thingfoaf:Agent"
>
> I don't see anything in the spec about how to split an
> attribute value into a list of curies.

Section 2.1 says that all the attributes are whitespace separated.


> Do I split by whitespace and then throw out the non-curies?
> (such as ***foaf:Agent, leaving just 1 CURIE, foaf:Person)

Yes -- see section 5.4.3.


> Or do I find all occurrences of CURIE syntax?
> i.e. 4 CURIES: foaf:Person, foaf:Agent, owl:Thingfoaf
> and Thingfoaf:Agent?

No. :)


> or all non-overlapping occurrences? i.e. 3

And no.


> What if foaf: is bound but owl: isn't?

In the current spec, the CURIE with 'owl' in is simply dropped.

In a new proposal, anything with an unbound prefix is *by definition*
not a CURIE, and is therefore a URI.


> Are there relevant test cases? I hope we come up with some.

There's a ton of them, although I don't recall one with two colons in
-- that looks a good candidate for a new one.

Regards,

Mark

Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: parsing lists of CURIES

Dan Connolly
On Thu, 2010-02-04 at 07:22 +0000, Mark Birbeck wrote:

> Hi Dan,
>
> > how many CURIEs in this attribute?
> >
> >  typeof="foaf:Person ***foaf:Agent 123412:abc owl:Thingfoaf:Agent"
> >
> > I don't see anything in the spec about how to split an
> > attribute value into a list of curies.
>
> Section 2.1 says that all the attributes are whitespace separated.

OK, but that section is informative.

I'm struggling with stuff that seems to be specified several
different times in the spec. Section 6 seems to re-specify
stuff from section 5 without any clear boundary. You're
pointing out that section 5.4.4 overlaps with section 2.1;
I don't see any cross-refences between them.

If something is going to appear in >1 place for convenience,
please have all but one of them refer to the main place.

>
> > Do I split by whitespace and then throw out the non-curies?
> > (such as ***foaf:Agent, leaving just 1 CURIE, foaf:Person)
>
> Yes -- see section 5.4.3.



--
Dan Connolly, W3C http://www.w3.org/People/Connolly/
gpg D3C2 887B 0F92 6005 C541  0875 0F91 96DE 6E52 C29E


Loading...