VOResource PR #1

Markus Demleitner msdemlei at ari.uni-heidelberg.de
Thu Dec 8 10:00:55 CET 2022


Dear Renaud,

On Wed, Dec 07, 2022 at 04:48:37PM +0100, Renaud Savalle wrote:
> There is a growing interest (from some EPN-TAP providers) to add
> contributor types (possibly controled by a Vocabulary) in the
> VOResource schema (similar to what can be done with DataCite - see
> line 7a of Table 4 in
> http://schema.datacite.org/meta/kernel-4.3/doc/DataCite-MetadataKernel_v4.3.pdf)
> 
> This is somewhat related to the note "Data Origin in the VO"
> currently being drafted by the DCP WG:
> https://github.com/gilleslandais/ivoa-dcp-data-origin
> 
> Now, it might be too early for VOResource 1.2, as I understand that
> more discussions are planned about "Data Origin in the VO" during
> the next Interop meeting in 2023.

My take on this would be that *if* we add contributor types, it
should almost certainly be a copy of what DataCite does, as everyone
would justifiably cross with us if we pulled up something that is
similar but subtly different.

Given that, it would be really simple to add a vr:Contributor
extending vr:ResourceName (contributor's current type) type
that has an contributorType attribute which would say: "See the
latest version of the DataCite metadata kernel for the values allowed
here".  That would suck a bit for validators that would need extra
code if they wanted to validate that, but as long as this is not used
operationally, I think we could live with having it not validated.

But that brings me to the central question: Why would we want this?
I've never been really clear on why one would want this contributor
thing in the Registry -- what sort of discovery would become possible
with it?  Or should that cover a non-discovery use case?

One scenario I could follow is when people generate DataCite metadata
from VOResource records (which is lossy but not hard; see
https://volute.g-vo.org/svn/trunk/projects/registry/dois) *and* want
these contributor types in the datacite records (where I can imagine
a few uses for them).  But does anyone actually plan to do this?

So... say if someone comes up with a concrete plan what to do with
contributorType in VOResource I'd not argue against having it in
version 1.2 -- but a "might be useful some day" wouldn't be strong
enough in my book to accept the somewhat odd construct referencing a
vocabulary written in an external and moving standard.

        -- Markus


More information about the datacp mailing list