table.load.cdf?

[Mark Taylor]

Hallo SAMPers.

There has been the suggestion to introduce a new MType table.load.cdf,
for exchanging tables in CDF format (http://cdf.gsfc.nasa.gov/).
This format is widely used for solar and space physics data as well 
as elsewhere.  It is typically, but not necessarily, used for
time series data.

It's clear what this MType would look like, namely the same as the
existing table.load.votable and table.load.fits MTypes currently
documented on the SampMTypes wiki page 
(http://wiki.ivoa.net/twiki/bin/view/IVOA/SampMTypes).
It is in experimental use in a test version of CDPP/AMDA, as
described here:

   http://cdpp1.cesr.fr/downloads/tutorials/import_CDF_in_AMDA_via_SAMP.pdf

On the one hand this is perfectly reasonable, but on the other hand
introducing many different table.load.* MTypes risks fragmenting
the community of SAMP applications.

At present the large majority of tools that exchange tables via SAMP
do it in the form of VOTables.  This means that SAMP table 
producer/consumer applications do not need to support a large 
range of data formats, since everybody sends and receives a single 
format, VOTable (although the MType table.load.fits exists, I don't 
think it's widely used in practice - does anybody know different?)

So one option would be to encourage the CDF users to convert their
tables to/from VOTable format for exchange via SAMP.
However this may not be appropriate for a community which is already
standardised on some other format and who may not be much interested
in most of the existing VOTable-friendly SAMP-aware tools.
Additionally, some differences between the data models imposed
by VOTable and CDF may mean that such transformation is problematic
(causing data or metadata to get lost or mangled).  In any case
there is no way to force people to use particular MTypes,
SAMP is intentionally open to people creating and using whatever
MTypes suit their purposes.

I think the best solution would be to adjust the existing text
on the SampMTypes wiki page, indicating that the table.load.votable
MType is the one generally recommended for interoperability, 
but that other format variants (table.load.fits, table.load.cdf, 
maybe new ones if there is a demand) are permitted for special
purposes where required.  It should be made clear that general 
purpose tools are not expected to implement them all.

If this suggestion is greeted on-list by the customary silence,
I will go ahead and adjust the text on the wiki page accordingly.
But if anyone has opinions on it, please speak up.

Mark

--
Mark Taylor   Astronomical Programmer   Physics, Bristol University, UK
m.b.taylor at bris.ac.uk +44-117-9288776  http://www.star.bris.ac.uk/~mbt/