inkdroid

Holy crap … now I need to listen to this. It’s so nice to know you’re not alone, and off on another planet.

I’m just now running across lingvoj.org, a linked-data application for languages created by Bernard Vatant. lingvoj basically mints URIs for languages (using the ISO-639-1 code) and when resolved (yay HTTP) nice human and machine readable descriptions about the language are returned. So for example the URI for Chinese is:

http://www.lingvoj.org/lang/zh

If you click on that link, your browser will display some HTML that describes the Chinese language, and if a client wants “application/rdf+xml” it’ll get back a nice chunk of rdf — all via a 303 redirect as it should be.

lingvoj is interesting for a few reasons:

• I work at the Library of Congress, who are the maintainers of iso639-2, and I know someone experimenting with a linked-data application for delivering it.
• I know software developers at LC and elsewhere who need access to this data in a predictable and explicit machine readable format, which lends itself to being updated (re-harvesting language URIs).
• lingvoj follows the 303 URIs forwarding to One Generic Document pattern, which is nice to see in practice. I also learned about the use of rdfs:isDefinedBy to assert (in this case) that a language is defined by the HTML representation for the language. Not sure how I missed that in the Cool URIs document before.
• There are owl:sameAs links between lingvoj and dbpedia and opencyc, which in turn are linked data, and allow an agent to walk outwards and discover more about a language. Maybe one day lingvoj could link to our ISO693-2 codelist at LC?
• lingvoj defines a vocabulary which includes a new OWL class Lingvo for languages, that happens to extend dcterms:LinguisticSystem.

It’s a lot o’ fun discovering this emerging, rich data-universe on the web. If you are the least bit curious take a look for yourself:

  curl --location --header "Accept: application/rdf+xml" http://www.lingvoj.org/lang/zh

Or better yet:

  rapper -o turtle http://lingvoj.org/lang/zh

Or if you are really adventurous grab the whole data set and put it into your triple-store-du-jour.

Recently there’s been a few discussions about persistent identifiers on the web: in particular one about the persistence of XRIs, and another about the use of HTTP URIs in semantic web applications like dbpedia.

As you probably know already, the w3c publicly recommended against the use of Extensible Resource Identifiers (XRI). The net effect of this was to derail the standardization of XRIs within OAISIS itself. Part of the process that Ray Denenberg (my colleague at the Library of Congress) helped kick off was a further discussion between XRI people and the w3c-tag about what XRI specifically provides that HTTP URIs do not. Recently that discussion hit a key point by Stuart Williams:

… the point that I’m trying to make is that the issue is with the social and administrative policies associated with the DNS system - and the solution is to establish a separate namespace outside the DNS system that has different social/adminsitrative policies (particularly wrt persistent name segments) that better suits the requirements of the XRI community. There is the question as to whether that alternate social/administrative system will endure into the long term such the the persistence intended guarantees endure… or not - however that will largely be determined by market forces (adoption) and ‘crudely’ the funding regime that enables the administrative structure of XRI to persist - and probably includes the use of IPRs to prevent duplicate/alternate root problems which we have seen in the DNS world.

It’ll be interesting to see the response. I basically have the same issue with DOIs and the Handle System that they depend on. Over at CrossTech Tony Hammond suggests that the Handle System would make RDF assertions such as those that involve DBPedia more persistent. But just how isn’t entirely clear to me. It seems that Handles like URLs are only persistent to the degree that they are maintained.

I’d love to see a use case from Tony that describes just how DOIs and the Handle System would provide more persistence than HTTP URLs in the context of RDF assertions involving dbpedia. As Stuart said eloquently in his email:

Again just seeking to understand - not to take a particular position

PS. Sorry if the blog post title is too cryptic, it’s Bowie’s “Five Years” which Tony’s post (perhaps intentionally) reminded me of :-)

Andy reminds me that a relatively simple idea (I think it was David’s at RepoCamp) for the OAI-ORE Challenge would be to create a tool that transformed OAI-ORE resource maps expressed as Atom into Google Site Maps. This would allow “repositories” that exposed their “objects” as resource maps, to easily be crawled by Google and others.

It would also be useful to demonstrate what value-add OAI-ORE resource maps give you: to answer the question of why not just generate the site map and be done with it. I think there definitely are advantages, such as being able to identify compound objects or aggregations of web resources, and then make assertions about them (a.k.a. attach metadata to them).