towards SED DM 1.0
chil at sai.msu.su
Fri Feb 11 07:23:53 PST 2005
After a discussion we had in Paris with Philippe Prugniel, I can propose
some steps to be made to evolve towards SED DM v.1
Isn't it a good idea to restrict and simplify the present data model and to
move all the not very clear and not clearly established features to the next
phase (SDM 2)? It will allow to have a version 1.0 ready very soon and to
start some other projects based on it (for example, 3D data model that I'm
Here is the list of what __might__ be done.
1) stop adding new features to the SED DM v.0.93beta (to freeze the feature
set): reject for a moment continuum-normalized spectra, non-calibrated
spectra, resolution (everything I proposed before) etc.
2) remove all the features that are not connected directly to the photometry,
for example for Flux.Value remove "Counts in spectral channel", "Count rate
in spectral channel", "Flux ratio of two spectra", and some others
3) remove all the features concerning spectral coordinate and flux
representation except, for example "frequency" in Hz and "F-nu" and/or
"vacuum wavelength" in Angstroms and "F-lambda", and to encourage data
providers to convert the data on their side.
4) Clarify the technical points: validate table "FITS keywords<->DM fields",
make a decision about VOTable serialization of arbitrary data model schema.
With best regards,
P.S.: Sorry, if you have this message more than once.
More information about the dm