Linking Open Data

This is a challenging project for Semantic Web technologies. Weaving together open public data, such as Wikipedia or Geonames, and public ontologies and vocabularies such as Wordnet, etc. Of course I had to be involved in that. But consensus will be hard to achieve. Initiators are folks from Leipzig and Berlin universities, involved in dbpedia, a project to RDF-ize Wikipedia content.

I've pushed the idea that linking concepts from different schemes should not be done on the basis of too strong ontological commitment, but of some kind of loose coupling using e.g., SKOS mapping vocabulary, and why not blank concepts. This proposal has not been well received, to say the least.

Chris Bizer says:

I'm in general against using bnodes for anything! They should be deleted from the RDF spec and they are especially harmful in a linked data context, where everything should be dereferencable.

And Richard Cyganiak adds:

Yes, sure ... I understand why you want to introduce the blank node. But I don't like it. Why do you generate data? You want it to be useful. How does this blank node increase its usefulness? It doesn't. It's just a fig leaf to cover up the fact that your model is just an approximation of the real world. But we know this. Every model is. And these "semantic rubber bands" don't change this fact -- they just make your data harder to work with. Be bold! People who want to re-use your data will learn to work around its quirks and idiosyncrasies. Dealing with the quirks is a part of re-using data, it always was, and it always will be.

So much for my hubjects ... but this is not the end of it. More to come.

Every subject is a blank node

Two discussion threads made me move a step forward towards a general theory of blank nodes. One thread I already mentioned is about languages. The other one was started here by John Black on Semantic Web list, about representation of concepts having contextual semantics, such as "I', "You", "Here" etc. Using a blank node to represent the context is the solution I propose today here.

More use cases for nondescript resources

nondescript is a funny word. Quite sure I did not catch its correct meaning when I first heard it. It has no real equivalent in French. Indescriptible is too strong, it refers to something too extraordinary to be described, whereas nondescript is actually too ordinary to deserve a proper description. Or is it that any description would miss the point? The free online dictionary says : Lacking distinctive qualities; having no individual character or form. Maybe in fact nondescripts remain such because those features which makes their individuality are too subtle to be captured, or this capture would be useless.

Anyway, let's take them as they are : no description. What does a nondescript look like in RDF? Well, easy, it's a resource with no description. No type, no identity, no properties, certainly not even worth a URI. A blank node with no property whatsoever. Rings a bell now? After the previous post, I've come out with more use cases for those nondescript resources, and how natural language uses them as binding points. Examples

John walks through the neighborhood where Ann is working.

:John :walksThrough _:n

:Ann :worksIn _:n

In The Rule of Four (currently reading) Vincent and Richard have different theses about Hypnerotomachia Poliphili. Paul prefers Richard's thesis.

:Vincent :hasThesis _:tV

:Richard :hasThesis _:tR

:Paul :prefers _:tR

Now the "Class or Concept" example comes naturally as just another case

a:Restaurant a owl:Class

b:Restaurant a skos:Concept

a:Restaurant :represents _:x

b:Restaurant :represents _:x

Cloud hidden, whereabouts unknown. (Similar stuff in the early '70s)

Blank Nodes continued

Steve Newcomb took the time to make an excellent comment on my previous post. Actually, this comment would have deserved a full post, and I remind my old friend that he has a permanent invitation to appear here as a contributor, and I would be very honored if he could join. That said, Steve's viewpoint is much closer to mine that he seems to think.

SRN : I insist that subjects do have identity, but only within contexts -- within universes of discourse.

I did not write something very different when I wrote that only representations can have identity. Maybe I should have put it slightly differently, and I'm sure Steve will agree with this other way to put it : Whatever the subject, it has neither absolute identity, nor absolute definition, nor absolute property of any kind, that would be valid in any context. Identity, and all properties bound to this identity, is always conferred through a representation, itself defined inside the context of some representation scheme, and making sense only in the framework of this scheme.

SRN: But I don't see how it's meaningful to say that a proxy is not a proxy for something in particular. By its very nature, a proxy is always a proxy for something in particular.

Well here I think I disagree with Steve, if "something" is to be understood as "some thing". My view on this has always been that, for all practical reasons, it's the first act of "proxyfication" which brings the subject into existence, as a subject of conversation. But this debate is not really important, and we can proceed from here to the notion of subjects as blank nodes, with or without agreement on the separate existence of subjects. The original point of the debate set by Alistair and Dan was to know how to express in an efficient and meaningful way the fact that two or more representations in different schemes are somehow proxies for the same subject. My point was that this could be captured by something quite similar to an RDF blank node, lets' call it a Subject Blank Node, bearing no absolute identity, and of which only properties could be : represented this way here, and that way there.

The way I see it, a Subject Blank Node would have no logical property per se. It would not be part of any representation scheme, but would provide a hub between various schemes. Such a hub would allow applications able to make sense of several representations schemes, each with their specific structure and logical rules, to aggregate information from different schemes, such as a Topic in a TM application, a class in an OWL ontology, a concept in a SKOS scheme, a category in dmoz, a page in Wikipedia, a term in Wordnet, or a picture by Van Gogh, or a Nocturne of Chopin.

Ontology Mapping, Ineffable Subjects and Blank Nodes

In this thread on SWAD forum, Alistair Miles and Dan Brickley re-activate an old issue : How do I express that resource X in representation scheme A (e.g. a SKOS concept scheme) and resource Y in representation scheme B (e.g. an OWL ontology) are somehow representations of the same (----) . After suggesting a suboptimal Topic Map solution I suddenly yesterday came out with the idea that in RDF, blank nodes could be a killer solution. Actually one can use blank nodes to aggregate various representations of whatever, keeping agnostic on what this whatever is. Using blank nodes to represent "ineffable subjects" is cool, since nobody is able to say anything directly about them (asserting name, type or any other property), since they have no URI. Put it together with recent debate on ISO SC34 mailing list about subject locators, and consider this provocative conclusion : RDF blank nodes are better than TM topics at representing subjects, since, and this is my last thought, subjects have no identity, only representations have one. Subjects have no identity, read no type, no property at all. Resources have identity (URIs), so the best attempt to indicate a subject is to gather various resources in a blank node, as so many fingers pointing towards the moon.

Remember in the Topic Maps book, I wrote about an empty subject indicator ...