Re: Grouing URIs - an RDF triple with multiple subjects

* Sandro Hawke <> [2006-05-08 14:02-0400]
> > The Web Content Label Incubator Activity [1] is to hold a face to face 
> > meeting in Edinburgh on Saturday 20 May. The major topic of conversation 
> > will be how one might define a group of URIs so that they, plural, can be 
> > the subject of an RDF triple.
> > 
> > Outside the XG there is interest in this from, among others, the ERT WG [2] 
> > (the folk behind EARL).
> > 
> > Some of you will have heard me talk about how we do this in RDF-CL [3] but 
> > we're trying to do a better job of it in the XG.
> > 
> > If the problem is one you'd like to crack as well, and you are in Edinburgh 
> > that Saturday afternoon, you'd be welcome to jon us and try to solve it. 
> > Please contact me directly. NB.we are all paying our own way - ??35.25 a head 
> > for the room [4] and sandwiches for a working lunch.
> I'm sorry I can't make it to the meeting, but I'm going to go ahead and
> offer a solution, since it seems so clear to me.  :-)
>    1.  You need a "uri" property, linking things to strings
>        which are unambiguous names (URIs) for them.
>    2.  You need a simple rule language.
>    3.  The rule language has to have simple/common string functions.
> Then you can say "everything which starts with is
> child-safe" like:
>     if 
>        x.uri.startsWith("")
>     then
>        icra:ChildSafe(x)
> I expect the RIF Working Group (for which I am the W3C staff contact) to
> provide something that will work for points #2 and #3.  I'm not sure the
> URI property is going to come from.

I'm very glad to see this conversation happening. I've always seen
RDF-CL as a stepping stone for PICS-like apps, with RIF being the
ultimate way of expressing the kinds of generalisations we want here.
I'm looking forward to test-driving RIF's language on these apps...

Re the URI property, perhaps this can be done in the XG. There might be 
a bit of care needed around canonicalisation, charset, considerations, 
also whether we use plain literals or datatyped, and if former, whether
we consider xml:lang relevant for identity reasoning. In other words,
we'd need a test suite.

For the string functions, SPARQL seems to have some useful bits:

Will some of these also be exposed in the RIF language?



Received on Monday, 8 May 2006 20:07:05 UTC