[data-cube] Comparing protocols

[Douglas Tody]

As it happens we had a discussion just like this here Friday, with the
CASA folks expressing concern about what would be required to get at
images/cubes via VO.  The main point made to us was that (for image
data) they only wanted to deal with a single service, providing
everything required to discover and access images within a single
service interface.  SIAV1 vs SIAV2 was also an issue, but could probably
be dealt with in the client side VO interface, that for basic access
could make the two look the same to client applications.

To me the main use for TAP/ObsTAP has always been the general archive
query interface.  If this is used to discover image data, then SAMP or
whatever is used to hand off a dataset reference to some image analysis
application that then uses SIAV2 to do the actual image data access.

 	- Doug


On Mon, 28 Oct 2013, Mike Fitzpatrick wrote:

> On Mon, Oct 28, 2013 at 5:20 PM, Robert J. Hanisch <hanisch at stsci.edu>wrote:
>
>
>> .......  In any case, it seems to be a wash, client-side.
>>
>
> Not really....if some providers do it one way, and others do it another
> way, and if the registry isn't clear on which is which and how to use them,
> it becomes a nightmare on the client-side.  Yes, it's easy if I want to
> build a client specifically for CADC or (DALServer-built) services, but
> throw in legacy SIAv1 services and those that aren't compliant in some way
> and a client to get an image from the VO in a general/easy way gets
> complicated quickly.
>
> My $0.02,
> -Mike
>