Converging the Data Models
bonnarel at alinda.u-strasbg.fr
Thu Sep 23 15:51:25 PDT 2004
About the new char schema provided by Jonathan.
As Mireille and I wrote to Jonathan in private the good thing with
that schema is that it is simple, human readable and that every element
has a specific name in characterization, although the xml types could
be eventually imported from Quantity or STC. This allows that charecterization
utypes keep simple and clearly identified as char, for eg use in SIA 1.1!
I have a few remarks on details, however.
1) sampling_precision ---> ScalarSize
If I understood well the evolution from Arnold's STC to JOntahan's
repeated in this char version, ScalarSize is coming from CoordSize and
the STC document explains that it is more for FOV, than pixel size ore
sampling. But there is a ScalarPixSize. We should use that one, should'nt we ?
2) Why does FrameType contain an element like this
and not (like in Quantity) <xsd:element name="type" type="cha:floatType"
which identify a specific dataType description in the Frame.
3) Why do we need cha:type to be defined as an extension of string?
Actually this is a question for Quantity? why do we need value cellType
to be an extension of string? instead of xsd:float,xsd:integer, ...
In the case of "char" Float would be sufficient,no?
If lo_include, hi_include, fill_factor are really important things
for the Interval like the filling factor, why is it attroibutes and not
elements? What is the genearal policy for elements versus attributes in
the IVOA schema? Isn't attribute for a "service" parameter?
By the way I have a general proposal which I know Jonathan doesn't like
very much, but it is really a "service attribute" proposal.
Elements in IVOA schema should allow a Utype (or DMHyperlink) attribute
and a ucd attribute.
Utypes would be for references to not totally consistent Data Models
(eg IVOA characterization references into Simple Spectrum, or different
Ucds attributes would be for the concept associated to the element name
This is NOT the same that the ucd element in FRame for example.
the ucd element in ScalarResolution will tell us which axis we are dealing
the ucd attribute for ScalarResolution will be the UCD1+ for Resolution
(sorry I don't remember!)
Regards to all
Francois Bonnarel Observatoire Astronomique de Strasbourg
CDS (Centre de donnees 11, rue de l'Universite
astronomiques de Strasbourg) F--67000 Strasbourg (France)
Tel: +33-(0)3 90 24 24 11 WWW: http://cdsweb.u-strasbg.fr/people/fb.html
Fax: +33-(0)3 90 24 24 25 E-mail: bonnarel at astro.u-strasbg.fr
More information about the dm