Lifting 1.1 limit in SCS?

[Mark Taylor]

On Thu, 18 Nov 2021, Markus Demleitner wrote:

> Hi Marco,
> 
> On Wed, Nov 17, 2021 at 01:19:59PM +0100, Molinaro, Marco wrote:
> > I took the liberty to draft the erratum:
> > 
> > https://wiki.ivoa.net/twiki/bin/view/IVOA/SCS-1_03-Err-2
> 
> Thanks for making that move.  I've expanded the erratum a bit, and
> I'd now be happy with it.

I prefer Marco's original

   "The XML content of the response must be compliant with the
    VOTable Recommendation."

over Markus's suggested:

   "Simple Cone Search services MUST return VOTable version 1 documents.
    This means that clients will have to handle the namespace URIs
    http://www.ivoa.net/xml/VOTable/v1.1,
    http://www.ivoa.net/xml/VOTable/v1.2, and
    http://www.ivoa.net/xml/VOTable/v1.3, where the last namespace
    will be used for all future VOTable 1 versions.
    For forward compatibility clients should ignore unknown elements
    and attributes in these documents as per the IVOA schema versioning
    policies."

Marco's version is simpler and preserves the parts of the intent of
the original that we want to preserve, and is probably what the
original authors would have written if they'd been thinking about
these things.  I don't think we need a lot of explanation to help
people out here (the impact assessment is that everything's working
OK anyway), we just need to unbend from the existing version restrictions.

I'm dubious also about Markus's mention of BINARY serialization,
only to point out that it's not relevant here:

   "We note that, in particular, the main interoperability problem even
    in the days of SCS 1.03, lack of support for BINARY-serialised tables,
    is orthogonal to the problem addressed here, as BINARY serialisation
    was already part of VOTable 1.1."

I would therefore prefer to delete the middle paragraph of the r2
impact assessment.
 
> > I'll try to work a bit better on the impact assessment
> > (if you already have some numbers there it would help).
> > I remember many services already returned VOTable-1.3.
> 
> Right -- the EuroVO validation service should give this number in one
> click, as I think "wrong VOTable version" is one specific criterion
> there.
> 
> But I have to admit I just spent five minutes to find these stats on
> registry.euro-vo.org or in the Ops sessions of the last three
> interops and failed.  Can anyone help out there?

I spent >5 minutes and failed too.  I suspect that this problem is
coded as one of the val_codes 2.2c-i, 2.2c-ii or 2.2c-ii,
as shown by e.g. ivo://archive.stsci.edu/catalogs/2mass
(http://gsss.stsci.edu/webservices/vo/ConeSearch.aspx?CAT=2MASS),
but I can't work out the mapping from those val_codes to a verbal
description of the issue being checked.

Mark

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