Dear all,
A quick followup to yesterday's telecon, with a proposed action on my part.
When I mentioned yesterday that we already have working tooling to use VOResource-style tabular metadata descriptions for guiding assisted query-builder tools, it was suggested that we were trying to drive the TAP standard based on a particular client implementation.
On the contrary, I raised the example to show that there already exists a *useful* IVOA VOResource format for describing tabular metadata, that is mature enough to have been integrated into real systems.
Speaking as a software engineer, I believe that the default condition would be for TAP to use the existing IVOA standard for expressing tabular metadata - there is no special justification required for doing this. On the other hand, for us to justify defining a new and competing standard, there would need to be a compelling argument as to why the existing standard is both currently inadequate, and not capable of being made adequate, for our purposes.
In order to anchor this debate to reality, I propose (unless anyone objects) to seek some advice from the registry group. I propose to ask their opinion as to the most appropriate existing VOReource tabular metadata description schema or schema fragment that we might use (together with any proposed developments to it).
Once we have this suggestion, we can then objectively consider whether this description format meets our requirements - and if it doesn't, our first course of action should be to consult the registry group and see whether an improved version can be produced. If this process fails to produce a usable standard, then of course we would be justified in defining a new one - but we haven't even begun the process yet.
Cheers,
Kona
"The nice thing about standards is that there are so many of them to choose from."