Re: Requirements for a possible "RDF 2.0"

On Thu, Jan 14, 2010 at 5:14 PM, Danny Ayers <danny.ayers@gmail.com> wrote:
> 2010/1/14 Hammond, Tony <t.hammond@nature.com>:
>>
>>> But please don't kill blank nodes !
>
> Sorry Chris on thread etiquette, but +1 to that.
At least, happy birthday Danny!

I've repeatedly tried to explain why not using bnodes where it would
be better use them is dangerous [1], if something might be dropped
(without loosing expressiveness) its names nodes, as Sandro already
suggested in 2001 [2] a uname property could be used instead.

In my experience explaining owl:sameAs is difficult, in fact it seems
to me an awkward way to express that a resource has multiple names, so
instead of relying on owl to express that multiple uri denote the same
resource it would be more straight forward to be able to specify the
different uris of a resource just using rdf. So maybe, instead of
dropping named nodes have nodes with 0 to n names and optionally a
literal value.

Ok, what scares me a bit are bnode properties. On one hand quite a bit
of code I've done would become more complex and significantly more
computationally expensive, on the other hand I don't really see the
use case for this, wanting to express that two pairs of resources
stand in the same relationship to each other without being able to
name this relationship seems quite seldom two me.

Cheers,
reto

1. http://lists.w3.org/Archives/Public/semantic-web/2008Jan/0118.html
2. http://www.w3.org/2001/12/uname/ and
http://lists.w3.org/Archives/Public/public-owl-dev/2007OctDec/0050.html

>
> There's a huge argument in favour of giving everything a (HTTP) URI,
> and as far as I can tell from the approach taken by Talis in the
> Platform stores, there's very little downside to this, the actual cost
> of minting & maintaining URIs is the least of our worries. Likewise
> timbl's Give Yourself a URI over everyone being blank nodes.
>
> But although getting rid of blank nodes may not be throwing the baby
> out with the bathwater, it's at least a limb or two. I'd point to the
> correspondence with variables in SPARQL, you can stick a ?x in the
> middle of a complex structure, even if you don't care about the
> value(s) of ?x itself. The related value(s) one can get through
> implication are worthwhile, even if the thing itself isn't available
> over HTTP. IMHO, €0.02.
>
> Love the Rumsfeld quote.
>
> Cheers,
> Danny.
>
>

Received on Thursday, 14 January 2010 21:38:05 UTC