Registries, IVO ids, and Data Set Identifiers
Ray Plante
rplante at poplar.ncsa.uiuc.edu
Mon Sep 22 22:34:58 PDT 2003
Thanks, Tom, for the clear discussion of the topics. I essentially agree
with all of it.
I just wanted to underline one of the the fundemental issues: journals
want to reference datasets; however, VO registries will not be registering
datasets. This is the impetus for a dataset identifer that is based on a
data collection identifier, which is registered. If we employ such a
technique, such as one of the ways Tom described, we can use VO Registries
to resolve datasets:
1. Drop dataset portion of dataset ID
2. Look up data colleciton in registry
3. Find the dataset resolving service that serves that collection
4. Use service to resolve dataset.
The only "change" needed in the IVOA ID WD is to reserve a character--# or
? (or both)--to indicated a "stop" character for the URI form of the ID:
everything including and after that character is not part of the standard,
registered IVOA ID.
On Mon, 22 Sep 2003, Tom McGlynn wrote:
> Something a little more in keeping with regular URLs might be
> to use a syntax like
> ivo://sa.rosat/x?set=rh3000001n00
> indicating that the id is a qualification of collection. Here a standard
> keyword like 'set' or 'id' would be chosen, but there would be natural
> path for expansion.
I'm okay with this if the ADEC gang is.
In general, unique IDs with multiple key-value arguments make it difficult
to compare IDs for equivalence (e.g. does order matter, are all args
required, etc.); but I don't think this is a requirement of the ADEC data
resolvers. If it is, then the # might be a better choice.
> One final choice might be to have the collection be everything before
> the final '/'.
> E.g., in ivo://sa.rosat/x/rh300001n00 the collection id would
> be ivo://sa.rosat/x.
I think it needs to be clear which part we expect to find in a registry.
It's possible that a journal may be based on an entire data collection
that is registered; in this case, the registered collection ID could used.
cheers,
Ray
More information about the registry
mailing list