validation of SIA 2

[François Bonnarel]

Hi Pierre,
     After the discussion on the list and our discussions in 
Stellenbosch I come back to summarize what shoukld be done from SIAV2 
spec point of view

Le 22/02/2016 19:15, Pierre Le Sidaner a écrit :
> Dear All
>
> I was working on the SIA2 validator on the last days
> I have a first version as a draft
> http://voparis-validator.obspm.fr
>
> I try it with the 3 examples taken from RFC Page 
> http://wiki.ivoa.net/twiki/bin/view/IVOA/Siav2RFC
> The SIA2 document is confusing because it refer to ObsCore without 
> telling exactly what is mandatory.
In SIAV2 context all the so called   "mandatory" fields of Obscore have 
to be implemented in the response. The optional may be added
  For a given service the SIAV2 response has always the same number of 
columns.
> Moreover I have question about SIA V2 and the implementations
>
>
>
> * could you please reference the schema in votable to validate it, 
> something like
> xsi:schemaLocation="http://www.ivoa.net/xml/VOTable/v1.3 
> http://www.ivoa.net/xml/VOTable/VOTable-1.3.xsd"  .
Yes we will add this in the spec examples. Probably an erratum is needed
>
> * for Utype example use an extension in the Utype name like 
> obscore:Char.SpatialAxis.Coverage.Location.Coord.Position2D.Value2.C1
> but as the Utype come from Characterisation, is it relevant to add 
> obscore, specially taking into account that Utype should be normalized.
In OBscore context, Characterisation is a Package. All Char utypes start 
xith "Char." All utypes are part of the Obscore name space
>
> * for request in the query. In the example inside the document, 
> REQUEST take the value query. Nothing in the documentation is related 
> to DALI for that. 
> http://dalservices.ivoa.net/sia_b?REQUEST=query&POS=CIRCLE 180.475 
> -18.70 0.01&BAND= 0.0008 0.0009&TIME= 55708 55710&COLLECTION=ALMA. It 
> should be more explicit in the documentation. And one ref 
> implementation does not support this parameter/value.
REQUEST=QUERY is mandatory when the service has several REQUEST 
capabilities. Currently SIAV2.0 has only one REQUEST option. IN that 
case DALI 1.1 says that "REQUEST=" is not supported (it is not usefull). 
The situation will be different with SIAV2.1
>
> * what about datatype and xtype are they both mandatory in the votable 
> output ? for s_region, s_fov ... in Obscore document there is both 
> adql:REGION and AstroCoordArea does both are acceptable. 
Yes the idea (in Obscore) was that these values of  TAP result tables 
may be used as queries  (an so it justifies     adql:REGION) but that 
stc:AstroCoordArea is more complete. Obscore 1.1 is currently in 
discussion on its  RFC page. So this is the right place to discuss this 
point. Probably Obscore text has to be clarified a bit.

> VOTable schema should include a list of xtype if it's possible
The xtypes are now properly defined in DALI 1.1
>
> * don't you see any problem as s_resolution have datatype=float and 
> xtype=adql:DOUBLE and sometime double in obscore doc.
Yes Obscore has to be fixed and the the examples too. WE will revisit this
>
> * what should I do with MAXREC, I mean what to test ?
>
That MAXREC is implemented and that the nombe of lines in the result is 
consistent. The SPEC explains what to do when MAXREC=0.

Cheers
François
>
>
> The 3 examples I took where
>
> http://www.cadc-ccda.hia-iha.nrc-cnrc.gc.ca/sia/v2query?
> it was my reference for test, if you try you must remove all the Inf
>
> casda implementation
> http://casda.csiro.au/casda_vo_tools/sia2/query
> does not support Request=query,
>
>
> VAO implementation
> http://vaosa-vm1.aoc.nrao.edu/ivoa-dal/siapv2-vao/sync?
> problem with DTD on votable
>
>
> Please give feedback about the validator, I am sure there is some 
> bogus left.
>
> Regards
> Pierre
>