Is anyone specifically answering Arnold's second post? It came after Pedro
gave out tasks. The issues were addressed in the session and presentation
that I attached, so I suppose someone could just answer with these
(interleaved in his large post):
RDBMSs implicitly convert strings to internal (datetime or timestamp) form
using a variety of techniques; ISO8601 format is an acceptable format
already. As with other string representations, it is a service capability
(possibly mandatory) to understand specific formats.
The defined geometry types are 2-D and the LONGITUDE/LATITUDE functions unnecessarily restrict it to spherical coordinates (see above for more about LONG/LAT). The PR was arrived at by balancing usefulness and complexity and the TEG feels we have found the best comprimise. We have considered future extensibility and found in some design experiments (and by looking at OpenGIS) that it is quite likely that to extend much beyond the current PR would not be possible. It is feasible to add 1-D regions (intervals) and 2-D ellipse in a way that is consistent with the PR. After that, one would have to break with SQL to add much value.
As described in the attached presentation (and above) STC can be used within the region construct; it is up to a service to specify that it understands STC and they can support as little or much as is feasible. The ADQL language itself remains independent of any specific string representation or version thereof. In the Interop discussions it was explicitly stated that we expect TAP services to use standards (for region and the coordinate system arguments). The PR will be elaborated to make this clear.
I will be out of email contact until the middle of next week, so if someone else wants to proof-read and post these 3 paragraphs into the RFC page, that would be great.
safe travels,
PS-Inaki: I mentioned to Pedro but I think only him: it would be good to track changes and have a new ref of document include change bars. I think it would be OK to accept the deletions and just show the additions.
-- Patrick Dowler Tel/Tél: (250) 363-6914 | fax/télécopieur: (250) 363-0045 Canadian Astronomy Data Centre | Centre canadien de donnees astronomiques National Research Council Canada | Conseil national de recherches Canada Government of Canada | Gouvernement du Canada 5071 West Saanich Road | 5071, chemin West Saanich Victoria, BC | Victoria (C.-B.)Received on 2008-05-23Z12:30:06