Some of the first things for discussion in Rwp02 as I see them are
i) the rough plan as outlines in the Initiate Discussion post.
ii) Scope definition, and
iii) Key Science cases.
Here I've jotted down some ideas on Scope Definition and how we might start on the Key Science cases, again flowing on from the London meeting discussion, and intended to motivate discussion here.
Some Ideas for Scope Definition:
In order to keep focused, the science- and use-cases are intended to be directly related to registry functions in the VO, rather than the VO as a whole. Or, at least the registry relevant aspects of general VO science cases should be foremost.
In considering registry functions, we should not get hung up on technical or implementation issues. For example we might simply refer to things like the "query format" without specifying what this format is, rather focusing on what resource and scientific information needs to be included.
Similarly we probably don't want to get bogged down specifying how a portal or interface may utiltise information returned from a registry, but at the same time not get too far removed from reality.
As part of the scope definition, it was also suggested
we try to decide on some definitions for the sake of
putting names to things like "resources", "query service" etc.
I'd suggest general adoption of the language and definitions
in "Resource and Service Metadata for the VO"
(http://www.ivoa.net/internal/IVOA/Registry19032003/ResourceServiceMetadataV6.pdf)
(although this may contradict Tony's meeting notes(?))
Issues of depth, complexity and granularity are bound to be central here, but we should keep in mind the general question as put by AL "what is the MINIMUM set of agreed standards we should agree on".
Key Science cases:
5 or 6 key science case envisaged to drive definition of
what needs to go into a registry. Let's start on one front
by taking the current science cases developed by the various groups,
and seeing what requirements these place on registries.
(NVO has of course defined some "registry use case scenarios"
already)
-Mark.