Re: SIA Evolution proposal

Just a very brief summary, a question to Pedro and Jesus (VILSPA), and, for "par condicio", a question to Thomas, Francois and Mireille (CDS), on the SIA evolution proposals:

http://www.ivoa.net/forum/dal/0145.htm (VILSPA) http://www.ivoa.net/forum/dal/0150.htm (CDS)

Both proposals manifest the same requirement: the need to provide enhanced metadata to a user of a S?A service (being it a SIA, SSA, etc).

The different is in the implementation, whereby:

• VILSPA uses VOTable "nested tables"
• CDS, which used VOTable "nested resources" in a previous version, in the last proposal simply uses a second GeneralFeatures resource parallel to the "main" resource (the one containing the S?A results). Such resource contains as many tables as needed to describe properties that otherwise would get repeated (eg filter descriptions, telescope location, etc). Very much like one would design his/her relational database.

   NOTE: In the CDS document that processed is labeled "factorisation",    which is probably coming from some french translation, I would call it "normalization"

   (from the relational word), but I'm italian ...

The VILSPA proposal will require a modification of the VOTable schema to allow TABLE as a possible datatype for a FIELD element.

CDS does not require any change, and indeed is already used by various data providers.

What is not clear to me, in the VILSPA proposal, how to avoid unnecessary repetitions.
The example by Pedro and Jesus shows a single record; but suppose that multiple records
(say 4) are returned, and suppose that 2 observations were taken in
energy band A and
2 observations were taken in energy band B. The same Energy_Band_Table will be
seen in two different records for both the A filter and the B cases.

In the CDS case that is done by putting all the Energy_Band (filter in the cds case) definitions
into a single filter table in the GeneralFeatures resource, and by referring to it.
(btw, the link should not be done on the filtername, but on a filter
identifier instead,
since the same filter name could be used by two different instruments
(see F555W for

both WFPC2 and ACS)).

Pedro or Jesus, could you please clarify that?

What is also not clear is how, with the new proposed CDS structure, one could describe multiple members of observations. In the previous IDHA version
that was achieved with "nested resources"; but now?

Francois et al., could you please clarify that?

A side comment more for the VOTable WG:
 I like the capability of nesting tables; that would allow a one to one mapping between
some special FITS file and VOTable. The HST/STIS extracted spectra are organised
in nested tables (a table cell contains a spectrum) indeed!

Thanks,

Alberto Received on 2004-04-15Z11:28:10