Next draft of VOTable in STC
seaman at noao.edu
Fri Jun 11 17:32:54 PDT 2010
I'm not sure NOAO as an institution can be said to have a single "eye" for standards, but your point is well taken. (NOAO, like other large astronomical institutions, might be more like the charming sea scallop, with dozens of cheerful blue eyes pointing in all directions: http://bit.ly/bbqIKi )
Steve is also too polite to point out that I've been one of the conspirators constructing an argument for more than a decade for strong adherence to the UTC standard :-)
That said, I phrased my email as a number of questions, not only because Socrates was a very wise fellow, but because I don't claim to have the answers. It is a very typical paradigm in astronomical software - let alone in the VO - for solutions to be proposed before the problem requirements are fully explored.
Casting the situation as the "impoverishment" of rich metadata is an implicit argument on behalf of a particular point of view. Ditto with focusing on "contractual specifications". Is the IVOA party to any contracts or MOUs?
I completely agree with the statement regarding the implications of our choices. We should not only mention the implications of our design choices in user documentation - it is even more important to recognize these implications ourselves. Implications have a way of cutting both ways, however.
My own point of view is to globally optimize the pragmatic benefit of IVOA standards on behalf of the very wide range of projects and stakeholders in the astronomical community. Sometimes this may hew closer to pedantic adherence to complex data models. Sometimes this may mean finding practical simplifications or even tolerable compromises with nonphysical models.
At any rate, it would be silly for the IVOA to beat ourselves up for failing to observe metadata standards that are widely flouted throughout the astronomical community. How to encourage astronomers to grapple with fully "correct" data models sounds like the makings for a wonderful ADASS BoF, or perhaps a topic for deliberation by IAU Commission 5. Is NOAO alone in bollixing FK5? Or is this widely the case?
I quoted "correct" in the previous paragraph because a standard that fully conforms to some physically correct model, but that is rarely implemented in practice, constitutes a solution that only Procrustes could love. The standard might not be arbitrary - but the result will be.
On Jun 11, 2010, at 12:41 PM, Steve Allen wrote:
> On Fri 2010-06-11T12:21:38 -0700, Rob Seaman hath writ:
>> How do non-IVOA projects handle FK5?
> Well, Rob, look at the log in your eye before the specks in others
> There is an FK5, with pointer to the original, but no explanatory
> comments indicating how the richness of the original metadata were
> impoverished to produce a simple consistency.
> I have had to deal with situations wherein the failure to appreciate
> the subtle meaning of the metadata has resulted in products which fail
> to meet contractual specifications.
> I have no objection to STC & VOTable deciding on a simpler model,
> but so doing has implications that deserve mention in user guides.
More information about the dm