Datalink vocabulary

[Norman Gray]

Pat, hello.

On 2014 Sep 23, at 17:46, Patrick Dowler <patrick.dowler at nrc-cnrc.gc.ca> wrote:

> On 23/09/14 03:01 AM, Norman Gray wrote:
>>   * extension#picture
>>     (an IVOA standardised 'extension' namespace): resolved to produce<http://www.ivoa.net/rdf/datalink/extension#picture>
> 
> I'm not so sure about this kind of usage. It means #foo is resolved vs the core vocab, but foo#bar is resolved against a "datalink" vocab in general... does the base URI for that even exist? Sure, there is a URL but no defined URI for "all datalink vocabularies"...

No, I don't think it would mean that.  Just as #image is resolved to produce <http://www.ivoa.net/rdf/datalink/core#image>, meaning 'image' in the namespace <http://www.ivoa.net/rdf/datalink/core>, the URI <http://www.ivoa.net/rdf/datalink/extension#picture> would represent concept 'extension' in the namespace <http://www.ivoa.net/rdf/datalink/extension>.  That's the URI that would have to be (sorry -- that it would be good practice to be) retrievable.  It would be tidy-minded to have <<http://www.ivoa.net/rdf/datalink/> also be retrievable, but not at all necessary.

It may or may not be worth mentioning the extension#picture mechanism in the document, but this scheme would make it possible to introduce such a thing in the future without changing the document at all.

By the way, it'd be a one-line change in the .htaccess file to make <http://www.ivoa.net/rdf/datalink/> redirect to the Datalink standards HTML.  Just say the word!

[Technical note: in XML, namespaced elements are a _pair_ of (namespace-uri, local-name); in the RDF 'namespacing' model, which is what we're essentially adopting here, all the properties and classes are named by absolute URIs, so that the 'namespacing' is really just an artefact of the process of resolving relative URIs against a base, and what 'namespacing' there is in this context is no more than DNS namespacing.  I mention this to point out the subtly different models, to avoid confusion -- needless to say, I don't believe there'd be any need to discuss this in any Datalink spec.]

> I would rather not mention this at all and just have hashtags->core for 1.0; then we can see how it goes with the simpler landscape.

Do you mean mentioning '#image' and 'http://example.org/foo#image', or not even mentioning the second?  I think the second is useful as a reassurance that individual projects have this freedom, and so don't have to argue much about the terms in the minimal core vocabulary.

See you,

Norman


-- 
Norman Gray  :  http://nxg.me.uk
SUPA School of Physics and Astronomy, University of Glasgow, UK