Comments on: "Bundling" instances of author names together without using owl:sameas

By: Co-referencing « LOCAH Project

Co-referencing « LOCAH Project — Wed, 13 Jul 2011 12:42:56 +0000

[...] Ben O’Steen, “Bundling” instances of author names together without using owl:sameAs” [...]

By: Hugh Glaser

Hugh Glaser — Tue, 07 Dec 2010 17:28:31 +0000

Hi Adrian,
Sorry to disturb you, even if we also interest you.
In fact the underlying CRS system doesn’t use owl:sameAs; see for example
http://southampton.rkbexplorer.com/crs/export/?uri=http://southampton.rkbexplorer.com/id/eprints-13916&format=rdfxml
But when we decided to gather everything into sameas.org, it became apparent that the service needed to go along with the sameAs usage, wrong as it may be. otherwise we would get involved with complex ontology discussions that would distract from the use of the service.
In those days, I think the whole thing was even less understood than it is now, and it was very hard to get people to accept the need for a coref-ontology, so they didn’t understand what the CRS software was doing.
In fact people can do a simply script to edit owl:sameAs to whatever their favourite predicate is, if they want.
One day I may add an argument to the interface so that you can tell sameAs.org what predicate(s) you want it back in.
Hi David,
I primarily think of bundles as Equivalence Classes, but sometimes refer to them as Closure :-).
Since equivalence predicates are reflexive, symmetric and transitive, they are equivalence relations (in the mathematical sense) and construct equivalence classes over the sets of URIs.
Closure is a related topic, concerned in part with whether, if you apply an operation to elements of a set (in this case the bundle, I think), you get an element of the set.
Congruence relations are equivalence relations (in the mathematical sense).
I’m sure someone else can do better on this if you want…
Best
Hugh

By: Adrian Pohl

Adrian Pohl — Tue, 07 Dec 2010 13:02:59 +0000

Hello Ben,

I am looking for a vocabulary to represent co-reference of bibliographic identifiers and am asking myself which approach fits best: - the coref-ontology, - your bundle ontology proposed in this post - or even owl:sameAs. (It is interesting - and disturbing - that Ian Millard and Hugh Glaser use owl:sameAs and not their bundling approach for the service sameas.org.) - or maybe even the Similarity Ontology?

I have the following questions to you: Why don't you (re)use the coref ontology in your approach although you explicitely built on the work by Hugh and Ian?What are the advantages of the bundle ontology comparing to the coref ontology?

Furthermore, I think it makes good sense relating the bundle ontology to OAI-ORE but shouldn't it read rdfs:subPropertyOf in lines 102 and 116 of the ontology, instead of rdfs:subClassOf?

Adrian

By: Name matching strategy using bibliographic data | Open Biblio (graphic) Projects

Wed, 01 Dec 2010 17:19:17 +0000

[...] – Bind the names together in whichever manner required. I would recommend creating Bundles as a record of a successful match, and an index or sparql-able service to allow the [...]

By: Disillusionment – 4Kids Entertainment, American Beauty in mind the whole OP (reprint, … | Home cooking

Sun, 28 Nov 2010 19:54:28 +0000

[...] “Bundling” instances of author names together without using owl … [...]

By: David F. Flanders

David F. Flanders — Tue, 23 Nov 2010 06:42:50 +0000

I’ve still yet to read ‘on co-reference and the semantic web’, but wondering if “congruent closure” (while not making a mathematical assertion necessarily) is similar to “equivalence relations”[1] in set theory (had to crack open an old set theory text to try and explore this, but couldn’t find any definitive answer)?

Also, great to see example code, could we maybe get an example query which demonstrates ability to get back a set of bundles?

Very interesting post.

[1]= http://en.wikipedia.org/wiki/Equivalence_relation

By: Tweets that mention “Bundling” instances of author names together without using owl:sameas | Open Biblio (graphic) Projects -- Topsy.com

Wed, 17 Nov 2010 14:43:21 +0000

[…] This post was mentioned on Twitter by Ben O'Steen, Julieta Lionetti. Julieta Lionetti said: RT @benosteen: Blog post on curating RDF, with a model for asserting that 2 or more URIs are the same… : http://bit.ly/8Xs8PQ @alfaqueque […]