SV: do we need it?

[Andrea Preite Martinez]

Quoting Tony Linde <Tony.Linde at leicester.ac.uk>:

> I'm wondering what astronomical terms would not be included in the SV?
>
The point of the draft under discussion is not to include all possible  
astronomical terms in the SV, but to be able to express any possible  
astronomical term, locally defined, in terms of standard SV tokens (or  
words, as we call them in UCD-ish).

Example 1. You are a data provider publishing in the VO your data on  
the optical V-band magnitudes of your sources. You locally call them  
v-mag. But v-mag is your dialect, so you register your data as  
providing phot.mag;em.opt.V, which is the standard VO way of saying  
optical magnitude in the V band.
Now, phot.mag;em.opt.V is an UCD, and UCDs are a VO standard that tell  
people how to form UCDs using the standard, maintained, list if  
UCD-words (or UCD-tokens!!). So, you are expressing your concept (your  
quantity) not looking for the UCD-equivalent of optical V-band  
magnitude in a list of UCDs (which does not exist!) but you are  
expressing your concept in terms of the standard UCD-words, in this  
example:  phot.mag and em.opt.V .

Example 2: suppose you want to register (I take the example of  
registering your data, just because this is the first step to move  
from local to VO-wide) your data on circumstellar matter around B  
stars. This is the sort of metadata information that you cannot  
express in terms of UCDs. We were not yet done with UCDs that we (all)  
realized the need to overcome this problem.
Is circumstellar matter around B stars an astronomical concept? Yes.  
Can you find it in the SV? No. The SV tells you how to combine the  
list of SV tokens in order to get something like:
diffuse;location.circumstellar;star.spectralType.B

Ex. 3: Or you can publish an alert on a GRB, using in the "what" (I  
suppose) field  the SV equivalence
time.variation.burst;ucd:em.gamma
(note here the use of the namespace ucd: to mark that em.gamma is a  
token in the UCD-words list) in order to be sure to be understood in  
the VO compliant community.

This approach is by far more flexible and practical than standardizing  
a list of all possible astronomical terms. I say so by personal  
experience. Last year I spent many weeks of tedious skimming through  
the way astronomers generate neologisms in astronomy. Take a look at  
Table 5 of the Note at  
http://ivoa.net/Documents/latest/AstroKeysTN.html .

There you can find a shortened list of well semantically defined  
terms, that are used by astronomers to mint a large variety of  
semantically different terms (or sub-terms, but this is not the point  
for the moment). And this is only the floating part of the iceberg! I  
also tested for variations around the term x-ray: I found more that  
700 different terms. I cannot tell you how many SV+UCD tokens would be  
necessary to express these 700 concepts, but I'll be happy to pay a  
round in Edinburgh if they were more than a few ten! We'll find a  
reason to drink anyway!

This said, I think the discussion going on (in particular on vocab  
formats) is extremely important, because the draft addresses the  
problem only at the, let?s say, the default, or minimal, level.
Being able to define higher levels could take us to a fascinating land indeed!

Andrea




===================================================================================
Andrea Preite Martinez                 andrea.preitemartinez at iasf-roma.inaf.it
IASF                                   Tel.IASF:+39.06.4993.4641
Via del Fosso del Cavaliere 100        Tel.CDS :+33.3.90242452
I-00133 Roma                           Cell.   :+39.320.43.15.383
                                        Skype   :andrea.preite.martinez
===================================================================================