SED Data Model version 0.92

[Jonathan McDowell]

Hi SSAP colleagues,
 
A brief catchup - sorry for the lack of progress recently.
I'll be out of email contact until Feb 7 and then
hope to get stuck in to SSAP work.

Alberto's message of Jan 25:
I agree we need to develop the Characterization model; but
this will still take some work, and I want to deploy this
interim spectrum model without waiting for it.

Randy's request for progress on common FITS file formats
for spectra: absolutely. The SSAP model FITS serialization
is a first try at this.

>     We just had our HST coordination meeting with CADC (Daniel Durand)
> and  ST-ECF (Alberto Micol). There was interest in reformatting the existing
> FITS spectral files to a common format for our final HST archive. Assuming
> you have not yet decided on what will be required for the SSAP, do you have
> any suggestions?  There was speculation that we may need at least 2 
> formats;
> one for single order spectra and one for multi order echelle. However,
> if we adopt a binary table format with one row per spectral order, both 
> could
> be easily handled.  Any thoughts? Any progress on the SSAP?

The one-row-per-order format fits fairly well with our SSAP data model.
Each order would be considered a separate segment in our model; 
we should add some extra metadata to indicate they're from the same
observation.

Randy's technical questions on the model:


>     First of all, I am confused why the document has changed from the 
> Simple
> Spectral Data Model to the SED Data Model when it will be used to 
> "represent
> SED, spectra and time series data".  Do you plan to write separate 
> documents for
> each data type?

No, the idea is that spectra and time series are just special cases of
an SED, so we don't do anything different for them.

>     Also, I am confused about the FITS serialization in section 8.1. Are 
> you requiring
> variable-length arrays and not allowing array fields within the binary 
> table itself?
> I would think not all data sets require variable-length arrays and the 
> FITS standard even
> recommends a simpler format when this feature is not necessary.

Hmm. Hadn't though of this, but sure - I guess fixed length arrays
should also be an option. But general readers will have to handle
the variable length case.

>     Finally, I think there are errors in the sample FITS header shown. 

Sorry, I'll fix this too. I made some changes and didn't do it consistently.
I already made some fixes in a draft 0.93 that I've shared with
Tamas, but I'll try and bring that up to proper standard by next week
and post it to the group.

  - Jonathan McDowell