[open-bibliography] URNs for National Bib Numbers

[Ross Singer]

On Thu, Dec 9, 2010 at 10:03 AM, Deliot, Corine <Corine.Deliot at bl.uk> wrote:

> In any case, I was thinking too narrowly. I don't know if NLS and NLW assign their own NBNs (I will try and find out) but in any case they could potentially do so, so I think we need to split the national namespace just in case.
>
I agree - they certainly publish their own bibliographies (at least
Scotland does, Wales doesn't seem to have anything analogous:
http://www.llgc.org.uk/index.php?id=readinglists), although there's no
indication of any sort of NBN.  Of course that doesn't mean there's
not some system they use internally (or that some number, based on
F-code, for instance, couldn't be adopted).

> What I'd like to know though is if the format Open Bibliography is choosing for the (BL) bnb will constrain the BL in the future if/when we decide to implement our NBNs as URNs. This is something we've been discussing.
>
Well, certainly it could force your hand a bit, depending on the
uptake.  In the absence of an established form, I think it's possible
that a de facto form could arise, if prevalent enough.

I think we're at the point where compiling all of the identifiers we
have about resources is a necessary step for disambiguation and these
identifiers are going to have to be formatted in unambiguous ways.

> What is the benefit for Open Bibliography? You've already created an identifier for the resource based on the BNB.

This is true, but it's non-standard.  So we (in this case, "we"
meaning "Talis") start using the Open Bibliography BNB URIs as the
identifiers for the BNB in Aspire or Prism (for example): is the BL
comfortable this (Open Bibliography) being the de facto identifier
assignment agency for the BNB?  If the Library of Congress were to
publish their catalog as linked data, how should they represent BNB
numbers?  What about the Open Library?  As linked library data gains
momentum, I think we're going to have to have identifiers like these
in place for people to correctly identify the resources in their
collections.

And, again, while URNs are less than ideal from a linked data
perspective, they are good for providing a consistent and unique
identifier, which is exactly what we need.

-Ross.