concept representation/displaying language?

[Francois Ochsenbein]

Mark, Carlos,

[Being a question related to VOTable, I'm posting it to the
VOTable working group]

>
>On Wed, 3 Jun 2009, Carlos Rodrigo Blanco wrote:
>
>>
>> As we are talking about vocabularies I've remembered a question that I posted
>> to the semantics group and I didn't get much help about...
>>
>> In some cases, we have to include in VOTables quantities (params, fields...)
>> that are not predefined in a protocol or data model or whatever (the point is:
>> no application will have a precoded way to
>> handle it)
>>
>> I would like to be able to tell the client (application) how that quantity
>> should be displayed (formated). To give a simple example, I would like to be
>> able to tell the application that this parameter that I call teff should be
>> writen as T_{eff} (latex notation). In this teff case it is not very
>> important, but in other cases it is... it's quite ugly to write
>> alpha_gamma_square when you can write \frac{\alpha}{\gamma}^2 (again latex
>> notation). And I think that applications should be as friendly to users as
>> posible.
>>
>> I can point to vocabularies telling the definition of quantities and relation
>> between quantities, I know, but it doesn't help.
>>
>> Does any one have any idea about such a thing? (You always can include a table
>> with fields saying how other fields must be represented using <!CDATA... and
>> so but, being it a solution, it seems to be just a patch as the application, a
>> priori, wouldn't know that it has to look for the information there)
>
>Carlos,
>
>there's a proposal to add a new attribute to VOTable FIELD and PARAM
>elements in VOTable 1.2; called something like xtype or extendeddatatype.
>Thus something like:
>
>   <PARAM datatype="char" arraysize="*" value="T_{\rm eff}"
>          extendeddatatype="latex"/>
>
>could work.  Of course the receiving application will need to understand
>the value "latex" of the extendeddatatype attribute.
>It may be required to add a namespacing prefix to the attribute
>value, and define its semantics somewhere else.
>
>I think this would be a reasonable solution to what you want,
>though note it's only going to work for FIELD or PARAM values,
>not, e.g. column names (FIELD or PARAM names).
>However the proposal is still to be drafted and accepted, so
>this won't work now, and depending on the form in which it
>makes it into the standard, may or may not in the future.
>
>Mark
>

I don't think the 'xtype' (or 'xdatatype', to be finalized)
is meant for this kind of info -- I would rather suggest 
the following possibilities:

A) either the name (name="\frac{\alpha}{\gamma}^2" is quite
   possible); 

B) or embed in the <DESCRIPTION>, (e.g.
   <DESCRIPTION>[latex_label=$\frac{\alpha}{\gamma}^2$] 
      Commputed value of...
   </DESCRIPTION>

C) or VOTable has to define another element or attribute to
   convey this info; if such details would have to be added,
   I would prefer to define of a new element like <TITLE> 
   or <LABEL> or <CAPTION> which could exist in addition to
   (or as a complement to) <DESCRIPTION>, something which 
   could look like
      <LABEL context="latex">\frac{\alpha}{\gamma}^2</LABEL>
   
Pros and cons:
A) is readily available, but may be in conflict with VOTable
   usage in some contexts (e.g. results of cone search)
B) is also readily available; but it is just a convention 
   which can't be enforced, and requires some extra parsing
   from applications to extract this information
C) requires a modification of the VOTable schema -- it could 
   be done if a strong requirement is expressed in the 
   immediate future...

Any other comment / suggestion ?

Cheers, francois

=======================================================================
Francois Ochsenbein    ------   Observatoire Astronomique de Strasbourg
   11, rue de l'Universite 67000 STRASBOURG  Phone: +33-(0)390 24 24 29
Email: francois at astro.u-strasbg.fr (France)    Fax: +33-(0)390 24 24 17
=======================================================================