[OAI-implementers] repository auto-discovery

[Herbert Van de Sompel]

Robert Forkel wrote:
> hi,
> i was wondering whether some sort of oai data provider auto-discovery
> should be part of the protocol? the way this is done for
> http://unapi.info/ seems pretty reasonable to me: just provide the
> base url of the data provider as href attribute of a link element and
> signal oai identifiers within the pages.
> does that sound reasonable? if so, should it be part of the protocol
> or rather an implementation guideline?
> regards,
> robert
> 

hi Robert,

(Good to hear from you on this list)

I have 2 comments to the above:

(1) As you probably know, the OAI has just launched a new, 2-year 
technical effort named ORE aimed at increasing cross-repository 
interoperability; see <http://www.openarchives.org/ore/>.  Within that 
context, I personally (and I speak for myself) have doubts that it is 
appropriate to release new Implementation Guidelines (let alone change 
the OAI-PMH) in the short term, i.e. before the new project has 
delivered insights regarding the specific problems that it will try and 
tackle and regarding the solutions to consider.

(2) A facet of the problem domain that unAPI covers has been explored in 
other contexts.  A generic formulation of that problem could be (if I 
dare) "How to obtain (in an interoperable manner) a manifest of all 
available Representations of a Resource" (to use W3C Web architecture 
language). The parties that I know have explored this problem area (see 
below) have all used different ways to frame the problem, different 
operating contexts to explore it, and different approaches to tackle it. 
  This problem, as well as all the existing ideas regarding ways to 
approach it will almost certainly be on the ORE discussion table.  And, 
obviously I hope that the ORE group can propose an informed and inspired 
resolution.

A few pointers to explorations of this problem domain in our community:

(*) unAPI: <http://unapi.info/>

(*) DLF-Aquifer Asset Actions Experiment: 
<http://dx.doi.org/10.1045/october2006-cole>

(*) the Pathways Obtain service hinted at in our recent D-Lib paper 
<http://www.dlib.org/dlib/october06/vandesompel/10vandesompel.html>

(*) The "obtain" service environment described in Jeroen Bekaert's Phd 
thesis <http://hdl.handle.net/1854/4833>.

But see also the recent 
<http://www.w3.org/QA/2006/10/missing_http_feature.html> as an 
illustration that this issue is not only on the minds of our community.

Greetings

herbert

-- 
Herbert Van de Sompel
Digital Library Research & Prototyping
Los Alamos National Laboratory, Research Library
http://public.lanl.gov/herbertv/
tel. +1 505 667 1267