too much complexity?

[Robert Hanisch]

I share Roy's concerns, and I think I've been consistent throughout this
project in urging folks to start simple, implement, and see what we've
missed or need to change rather than trying to one grand all-encompassing
design from the start.  By the time we get something like that right, much
of the supporting technology will have changed and we will be obsolete
before we have anything to show.  A small core done well seems to me better
than a complete system that many will not understand, or not take the time
to learn.  One maxim of the VO is to keep the entry cost low.

This is not meant to undermine Ray and Tony and others who are thinking hard
about the registry capabilities and implementation.  But Roy's cautions are
important, and we should seriously consider limiting scope for now.  The
danger, of course, is that we build and then fail to update in a timely way,
locking ourselves into an incomplete system.  Let's recognize this risk,
pledge to remain open and flexible, but try to reach initial closure by the
time of ADASS.

> (4) Why are there suddenly five kinds of linking relationship? If simple
> "citation" is good enough for the Journals, why is it not good enough for
> VO? Half the people filling in these forms will do nothing in response to
a
> complicated question -- and so we lose metadata -- but they will recognize
> and respond to the word "citations".

Following up this point briefly... I don't think the registry metadata
should even attempt to track citations TO a Resource.  This is dynamic and
open-ended, and folks like our colleagues at ADS and CDS are much better
equipped to take on this role.  They provide Resources to find citations.
The DC metadata element Source is much simpler and is sufficient to
establish a link to the literature associated with a Resource, i.e., showing
parentage but not offspring.

> (6) How many registry entries will there be for Vizier? If it is only one,
> then I suspect few people will be interested in the registry. If it is one
> for each of the 5,000 catalogs, then how many fields in VOResource will be
> filled in for each?

I think the registry should be able to handle each of the catalogs.  If
Resources are too coarse-grained we will have gained little over a search
done in Google.  If they are too fine-grained, e.g., to the dataset or
object level, we will have registries with billions of entries and they will
be unmanageable, or will require the kind of curation associated with huge
surveys.  My experience in hand-entering Vizier catalogs into the NVO
prototype registry suggests that the whole thing could be automated, e.g.,
by parsing the GLU dictionary and creating a mapping to the Resource

Cheers,
Bob