ADEC and VO data registration

[Alberto Accomazzi]

Hi Roy,

still catching up with VO-related emails...

Roy Williams wrote:
> (1) Your list of datacenters (*) looks to me like a list of what VO is
> calling publishing authorities. This is a little more general than just
> datacenters -- it would allow non-telescope data to be published (eg theory,
> eg derived or federated data). The VO registry scheme is also based on
> publishing authorities. Therefore to me, the "astronomy data centers and
> archives" that you work with are examples of "publishing authorities".

agree.

> (2) In this case the questions are who can be a publishing authority, and
> what is their responsibility if they are. Equivalently, we can ask how the
> list of publishing authorites is extended. What is the proposed review
> process that allows flexibility and novelty, but keeps out maniacs and
> nutters?

I don't have a good answer to your question, except to note that this 
problem is the same that we (ADS) face on a daily basis as nutters 
contact us asking to create entries in our bibliographic database to 
include their latest "paper" refuting the theory of relativity.  The 
"work" they want included in ADS is invariably something they have put 
up on their website.

After the recent rounds of discussions about adopting a VO-based 
approach for the verification and resolution of identifiers, I think the 
immediate problem is shifted onto the publisher's and VO-maintainer's 
lap.  The paper being published will be refereed, which means the 
information contained in it (including the links to datasets) will be 
under the scrutiny of referees.  The links to the actual data will be 
verified by tools that look up the datacollections in the VO registry, 
taking the view that if it's in the registry and it has been marked as a 
permanent identifier, it's good enough for us to link to.

I don't know if this has been debated before, but in my mind there will 
need to be an entity or consortium that decides what is accepted as part 
of a "master" VO registry, so that the technologically savvy nutters are 
still kept out.

> (3) Your validation service for datasets seems to return something simple
> ("yes", "no", or "bad syntax"). Is that true? However, the VO registry will
> return a sophisticated piece of metadata (VOResource.xsd). This is because
> of our somewhat different purposes: in VO we want to find and use the data
> and its provenance, but your registry is simply asking if the dataset
> exists. I wonder how a VO-registered dataset can make its way to your
> registry?

As you have seen from previous emails, in the current proposal the 
verification and resolution of a dataset ID is performed outside of the 
registry by services that are listed in the registry.

> (4) Sematically, the identifiers look similar, and are similar to the ISBN
> number for books. A publishing authority followed by an ID that is local to
> that authority. Would it be possible to iron out some of the syntactic
> differences between identifiers -- the use of colon, slash, etc etc is
> different? How concrete is your identifier syntax?

I always thought of DOIs rather than ISBN numbers when thinking of the 
identifiers, but either analogy seems valid.  One thing that is 
different is that we may want to assign a dataset identifier to a 
_collection_ of datasets.  For instance, a paper discussing a survey of 
HST data may want to refer to the collection of the datasets as a whole 
rather than 200 different identifiers.  So I think the DOI analogy is 
better: no all identifiers are "equal" (they're not all books), but they 
can all be tagged with a DOI.  Unlike DOIs, however, we do not require 
that each one of them be registered with a central authority.

In terms of finding a common syntax, I think what Ray proposed a couple 
of days ago sums things up well.


-- Alberto

****************************************************************************
Alberto Accomazzi
NASA Astrophysics Data System                     http://adswww.harvard.edu
Harvard-Smithsonian Center for Astrophysics      http://cfa-www.harvard.edu
60 Garden Street, MS 31, Cambridge, MA 02138 USA
****************************************************************************