concept representation/displaying language?

[Norman Gray]

Carlos, hello.

On 2009 Jun 5, at 16:57, Carlos Rodrigo Blanco wrote:

> I think that the nicest way for something like this would be to do  
> something similar to what is done for ontologies, vocabularies,  
> namespaces and so. I'm not an expert on those things, so please  
> correct me where I'm wrong.
>
> I would have some xfd (Xml Formatting Definition or whatever, I've  
> just invented the term) document following a TBD specification. In  
> that document I would have:
>
> <concept name=teff>
>  <label context=latex>T\_{eff}</label>
>  <desc context=latex>....</desc>
>  <label context=html>T<sub>eff</sub></label>
> ...
> </concept>
>
> or something similar (I just want to suggest the main idea)
>
> And, in the votable tag I would add the address of this document in  
> the usual fashion, somethinkg like <votable xfd:http:...> (or even a  
> INFO element could be given in the votable telling the address of  
> this document?)
>
> In that way, somebody can download the simple votable and, if he is  
> interested in displaying, he can download the xfd document and do  
> the matching. I don't know if the matching should be done by the  
> parameter name or by some utype or whatever (how is it done in SKOS  
> vocabularies and those things? I actually think that this thing that  
> I'm asking for could be an addition to vocabularies: not only giving  
> "definitions" and relations, but only suggesting the correct  
> displaying in several contexts)

In SKOS, and in ontologies, and other things in this domain, all the  
concepts and classes are named by URIs, and this (a) provides a unique  
name to refer to the concept, as well as (b) a place from which to  
retrieve the object.

Thus <http://www.ivoa.net/rdf/Vocabularies/AAkeys#Astronmetry> is the  
name for 'Astrometry' in the Proposed Recommendation IVOA SKOS  
vocabulary for the journal keywords.  If you download that URL in a  
browser, you get a human-readable explanation of the concept, and if  
you download it asking for RDF:

% curl -H accept:text/turtle -I http://www.ivoa.net/rdf/Vocabularies/AAkeys
HTTP/1.1 303 See Other
Date: Fri, 05 Jun 2009 15:18:20 GMT
Server: Apache/2.2.3 (Red Hat) DAV/2
Location: http://www.ivoa.net/rdf/Vocabularies/vocabularies-20081104/AAkeys/AAkeys.ttl
Connection: close
Content-Type: text/html; charset=iso-8859-1

% curl -H accept:text/turtle http://www.ivoa.net/rdf/Vocabularies/vocabularies-20081104/AAkeys/AAkeys.ttl
@base <http://www.ivoa.net/rdf/Vocabularies/AAkeys> .
@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
@prefix dc: <http://purl.org/dc/elements/1.1/> .
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
@prefix : <http://www.w3.org/2008/05/skos#> .

<>
     dc:created "2008-11-04" ;
     dc:title "Vocabulary for Astronomy & Astrophysics Journal  
keywords (Version 1.16)"@en ;
     a :ConceptScheme ;
     :hasTopConcept <#AstrometryAndCelestialMechanics>,  
<#AstronomicalDataBases>,  
<#AstronomicalInstrumentationMethodsAndTechniques>, <#Cosmology>,  
<#Galaxies>, <#InterstellarMediumNebulae>,  
<#PhysicalDataAndProcesses>, <#SolarSystem>,  
<#SourcesAsAFunctionOfWavelength>, <#Stars>, <#TheGalaxy>, <#TheSun> .

<#AccelerationOfParticles>
     a :Concept ;
     :broader <#PhysicalDataAndProcesses> ;
     :inScheme <> ;
     :prefLabel "Acceleration of particles"@en .

[...]
%

...you get a redirection to a file which explains what these concepts  
are in SKOS terms (the redirection seems roundabout, but is a best  
practice, for reasons which I can expand on if given the slightest  
encouragement).  (and don't worry about the syntax -- parsers exist  
and are small)

If utypes were URIs, you could do the same with them.  In such a case,  
it is both architecturally principled, and feasible in practice, to  
attach other 'remarks', not in the SKOS namespace, to the utypes.  It  
is also architecturally sound to make these annotations, referring to  
the same URIs, behind a different retrievable URI, in other words as a  
more or less standardised 'third-party' comment, if it were felt that  
including these in the utype definition would be too heavyweight.

This sounds very similar to what you've described, and is W3C- 
standardised (not the formatting vocabulary as such, but the mechanism).

Best wishes,

Norman


-- 
Norman Gray  :  http://nxg.me.uk
Dept Physics and Astronomy, University of Leicester