About the "DAL cube caravane" status and delays

[François Bonnarel]

Hi Dal WG members (and DM friends)

This email to try to clarify where we are for "Cube access" priority.

The deliveral due were the following for the cube "caravane"
DataLink 1.0, SIAV2.0 and AccessData 1.0 initially. After that came 
Obscore 1.1 as an additional effort.

DataLink is hopefully close to recommendation.
ObsCore 1.1 is probably in good shape.

SIAV2.0 : is delayed. We got a lot of comments in the TCG period and we 
are still lacking a client implementation. (I am personnaly working on 
this one currently because I think Aladin is a very good candidate for 
that). We are  also working on the examples and modifications of the 
spec and will propose something very soon.

AccessData : we have some kind of compromise discussed in Calgary but 
progresses are delayed by the situation in SIAV2.0. I am confident we 
will be able to have a new draft version for Sesto.

This makes a landscape with significant delay in the agenda. But we hope 
to make significant progress before Sesto and in Sesto.

During the discussion of Obscore 1.1, two topics raised which must be 
adressed in some way without too much interference with main agenda: DAL 
access to WCS and redshift axis charactisation in the response.  There 
is no DAL document on this but it can be discussed in a dedicated 
session is Sesto.

* The initial scheduling was that everything like coordinate mapping, 
WCS , etc.. was to be done in SIAV2.1 (so to be started after completion 
of 2.0) and was the so called GetMetada. Should be based on 
serialization of Image DM and coordinate transformation.
    The discussion beside ObsCore 1.1 showed there were pragmatic 
approaches to WCD DAL access based on prototypes. Some people need to 
develop things along these lines rapidly.

     We could imagine having several level of GetMetadata : naive, 
usual, extended. Naive based on blob WCS retrieval, Usual based on 
Tabular information like TAP  and extended based on some serialisation 
of the full ImageDM (xml, VO-DML,  or whatever....)
     So this is a topic to be discussed with DM people because it may 
need some kind of Image DM serialisation.

  * relationship between redshift axis and spectral axis is a complex 
thing and can probably only solved in the frame of the ImageDM ... At 
the Datadiscovery level of SIAV2.0 and Obscore 1.1 we probably have to 
keep simple. This also can be jointly discussed with DAL and DM people.

Best regards
François