[Metadata] DOI discussion paper

[Sally Morris <smorris@wiley.co.uk>]

Here is a potted version of my comments to Norman on his paper (note that
this was based on the MK1 version).   Apologies if some of this seems like
old hat - I realise discussion has moved forward while I was away!

Sally

1 - Scope
I do feel that a working group on this needs to be set up with some urgency.

As you know, it is my view that we should see the DOI as a 'digital
identifier of objects'.   I think that to exclude non-digital objects
(Godfrey's 'packages') unduly limits the potential benefits of the DOI (e.g.
for use in subsidiary rights trading).

Perhaps we should talk about 'creations' rather than 'objects', since
Godfrey's classification limits the latter to digital objects.

2 - Terminology
I agree with you that we need to make certain distinctions in what we are
talking about when we say 'DOI'.   However, I suspect you fall right back
into the same trap with your use of the phrase 'the DOI system' to describe
the loose conjunction of identifier, resolver, services/applications and
(possibly - see below) metadata!

3 - 'The DOI metadata'
This concept really bothers me.   I do think the entire metadata working
group needs to be actively involved in discussion of the whole idea.

I'm not even sure whether you mean:
        a)      All metadata associated with the object
        b)      That metadata which is needed to carry out transactions
(since 'get' is not appropriate in the e-commerce environment)
        c)      A minimal subset of one or other of the above, held
centrally (NB before establishing this we need to be absolutely clear about
its purpose;  Helen Atkins' subset is probably the elegant minimum for
reference linking, but for providing identification information in response
to a presented DOI I think you would need much less)

4 - Operations and (Third-Party) Services
I think these need to be clearly distinguished in your paper.   Forgive me
if my lack of techy know-how has got me confused;  but what I'm thinking of
is the difference between something which is automatically 'called' by using
the appropriate instruction (suffix, data type or whatever) and something
which is created by a third party such as Copyright Direct (or any other
intermediary).


Section 2.2 - note that simple 'get' is not applicable to any object which
is only available under terms and conditions of e-commerce;  last sentence
suggests that it is.

p4, first bullet - the idea of standardised services fills me with horror!
Standardised operations, possibly;  but the whole point of interoperable
standards and language is that any third party can create and offer services.

Bottom of p4 - the MWG is trying to do exactly this.   May be appropriate to
say so?

4.1, para 2 - packages have metadata too, surely?   Whether implicit or
explicit...

Para 3 - 'given this data, give me the DOI' only works if the object HAS a
DOI!   Further, I'm not yet convinced that the IDF has to perform this
function.   That's not quite how it works for other identifiers (e.g. ISBN),
is it?

I really don't see why, for other than internal IDF purposes, the name of
the assigner is required.

Note that the precise delineation of 'bare bones' metadata is
medium-dependent (e.g. journals are different from books...)

Last para - I'd go further;  metadata ARE the essential building blocks of
services, whether in the 'operations' sense in which I think you've used it
here, or third-party services.

4.4, para 1 - I suspect many of your readers will know even less about RDF
than most of us did before the NY workshop.   A sentence or two here (in
addition to the full reference) might help.

Last sentence - the problem is that there aren't enough existing standards;
this is fundamental to what the MWG is trying to enable.

4.6 - also necessary to mention that some metadata changes over time (e.g.
ownership;  terms and conditions;  etc)



Sally Morris
Director of Copyright and Licensing
John Wiley & Sons Ltd, Baffins Lane, Chichester, West Sussex, PO19 1UD
Tel:    01243 770365     Fax:    01243 770429      Email:  smorris@wiley.co.uk


------------------------------------------------------
Metadata maillist  -  Metadata@doi.org
http://www.doi.org/mailman/listinfo/metadata