Dear all,
I have been quiet reading the different issues concerning our ADQL doc, nearly finished.
I would like to remind you that we should be in the period of answering the RFC comments, rather than introducing ourselves new functionalities (that ought to have been discussed before, within the TEG) or significative changes to the document.
Nevertheless, some small or agreed changes (during the Trieste meeting) are necessary, and therefore the following is my proposal to close the ADQL periple towards PR:
1.- We leave ELLIPSE out of this version of the document, and start
thinking on whether it is well defined or not to eventually include it
in a future version.
The fact that different well experienced people have different views
on this very issue makes me skeptic on its easiness, and rather looks
to me like the inclusion of the ELLIPSE function needs further
thinking. The fact that we are issuing this version of the ADQL as PR
does not mean that the ADQL will be frozen forever. We have agreed
that this document will evolve in the future, and things might be
added to it, but they will have to be discussed properly within the
team and well researched and agreed.
2.- We adopt DEGREES (and square degrees) as units. This is a non- issue in my understanding, as conversions among them are trivial. It is just a matter of decision and agreement, and I believe we can all live with degrees, and will take the necessary steps to convert to whichever units we'd have our DB data in.
3.- We include BOX as per the latest inputs from Jeff. They have been agreed by Pat followed by Alex, so I give the latest definition as valid
4.- I have not understood the change of system_defined_function to geometry_function. In my understanding system_defined_function is more general than geometry_function and could allocate other types of system defined functions to be included in the future if necessary that not necessarily would have to be geometry functions. Therefore, unless there is a good reason against it, I call for not modifying the BNF on anything that has not been agreed, and leave it in this case as system_defined_function. This is not a criticism to Jeff's work, which is excellent and incredibly helpful for the ADQL doc, but just a reminder that we are only making changes in response to the RFC, and that we should refrain from introducing changes, despite how small, that could create more questions that the answers they provide with. As soon as the changes that reflect the comments from the RFC are included in the document (together with the agreements from the Trieste meeting), the document shall be ready for PR; thus, no changes other than the agreed ones should be put in.
So, to summarise, I call for your agreement or otherwise on the following plan:
I will send _tomorrow_ a mail to the community saying that the RFC period is closed, then Inaki will edit the doc to include all the final issues, and once it is ready we'll distribute it to the TEG, and -if accepted- will pass it to the IVOA TCG for final examination.
That's all. Please send me you agreement (hopefully) or otherwise, with comments if any.
Thanks.
Cheers,
P.
European Space Agency (ESA)
European Space Astronomy Centre (ESAC)
Science Operations Department (SCI-O)
Science Archives and Computer Support Engineering Unit (SCI-OE)
e-mail: Pedro.Osuna-at-esa.int
Tel + 34 91 813 13 14 Fax: +34 91 813 11 72