Difference in polygon's between SIA2 and ADQL

[Patrick Dowler]

I general we came to the conclusion that trying to support STC in ADQL/TAP
was a bad idea because it caused a lot of other pain. In fact, the STC
usage in TAP is at best "try this out" because it is in a non-normative
section of the document in TAP-1.0 and is being removed in favour of
simpler DALI geometry support. (I'm not including complete metadata about
columns in the VOTable header part -- that is still a good tuing to do and
hopefully vo-dml will formalise that.)

That said, simpler meant removing the coordinate system metadata and
starting with a few commonly needed shapes... I'm not too picky about which
definition of polygon we go with or think is best and it probably isn't too
late to argue for an erratum if you think the inconsistency between SIAv2
"polygon" and STC is a bug in the spec. Or, one could just say that the
SAIv2 polygon as used in POS is an idiosyncratic case and instead target
the polygon as defined in WD-DALI-1.1 to be the one that is consistent with
STC and/or able to cover more than half the sphere... that is a WD that
will effect SODA-1.0, TAP-1.1, etc.

Pat




On Wed, Feb 10, 2016 at 1:46 PM, Walter Landry <wlandry at caltech.edu> wrote:

> Patrick Dowler <pdowler.cadc at gmail.com> wrote:
> > I assume "complicates our implementation" means "have to compute the
> > winding direction and maybe reverse vertex order"?
>
> The trouble is just that it is different.  I would like to use the
> same underlying implementation for TAP and SIA.  Also, the SIA version
> makes it impossible to select more than half the sphere.  So a search
> doing something like excluding the galactic plane would be impossible.
>
> Cheers,
> Walter Landry
>



-- 
Patrick Dowler
Canadian Astronomy Data Centre
Victoria, BC, Canada
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ivoa.net/pipermail/dal/attachments/20160210/dad7ee92/attachment.html>