Re: Piggy-Bank feedback

From: Frank Manola <>
Date: Fri, 11 Feb 2005 17:13:18 -0500

A further data point on this thread about dealing with untyped nodes.
I've been going through the exercise of looking at various examples from
the RDF Primer (as individual files) in PiggyBank. One situation where
not dealing with untyped nodes creates a bit of a problem is if anyone
wants to use parseType="Collection" (example 17), since the generated
blank nodes are untyped, and as a result the structure of the list
doesn't show up. I suppose you could always create the lists in
"longhand" (example 18) and define types for the intermediate nodes
explicitly, but that's a bit of pain.


PS: I'd note in this connection that under the RDFS semantic
conditions, these blank nodes *do* have a type, namely rdf:List. Just
another complication.

Eric Miller wrote:
> On Feb 2, 2005, at 3:44 PM, Stefano Mazzocchi wrote:
>> DuCharme, Bob (LNG-CHO) wrote:
>>> I think that dealing with untyped nodes in proper RDF should be a
>>> higher
>>> priority than dealing with non-RDF, because Piggy-Bank already goes
>>> so far
>>> in breaking the chicken/egg cycle of
>>> not-enough-RDF-data-out-there/no-good-apps-to-take-advantage-of-RDF-
>>> data. If
>>> can find triples in an RDF file, then
>>> Piggy-Bank should be able to do something with them.
>> I don't think it's hard to stop PB from harvesting untyped nodes, the
>> problem is what do you do with them, I mean, in what category you
>> place them in!
> yep.
>> We might have a "misc" category where all the untyped stuff ends up
>> being.... but honestly, I think it's a lot better, at this stage, to
>> kinda "push" people to type their data rather than adjusting PB to
>> digest their unflavored one ;-)
> I'd suggest that a 'misc' (or something) category eventually makes
> sense. As one navigates, and runs across more RDF data, the type of
> information associated with a resource may be merged or inferred. At
> that point the resources would move from the 'misc' category into
> something(s) else.
> Further, I think it would also eventually be handy to allow the end
> user to associate one or more type(s) with these at any time. The
> ability to add comments now to this is a good start, but there is more
> than can, should be done ... plus enabling user / RDF editing is on the
> list of Simile deliverables :)
> That being said, like you mentioned I like the idea of using this
> 'feature' of not rendering untyped nodes as a means for getting people
> to type their work. Focusing work on fresnel and RDF editing, however,
> should enable the end user to do (smarter) things with unknown /
> untyped data. Make sense?
> --
> eric miller
> semantic web activity lead
> w3c world wide web consortium
Received on Fri Feb 11 2005 - 22:04:44 EST

This archive was generated by hypermail 2.3.0 : Thu Aug 09 2012 - 16:39:17 EDT