1	A practical approach to resolution of identifiers Experience from the DiVA project Eva Müller (eva.muller@ub.uu.se) Uppsala University Library Sweden
2	DiVA (Academic Archive Online) Project started 2000 at Uppsala University, Sweden Objectives of the DiVA Project: Develop technical solutions & workflows supporting full text publishing, storage and dissemination of university research Explore ways to ensure future access, ability to render and understand digital objects in the archive 2006 – cooperation within 17 universities in Sweden(15), Denmark (1) and Norway (1) Cooperation with the National Library from the very beginning
3	Long term access How can we ensure access to digital objects we produce locally? A stable point of reference (persistent identifier) Storage in several locations (cooperation with trusted organizations) Why to implement IDs and ID resolution? Stable point of reference Helping to solve the ” 404 not found –message” problem Making possible to move services and objects without risk of “broken links”
4	Some requirements for IDs and their resolution For the DiVA project we need an ID scheme which allows us Easy, reliable and cost-effective maintenance and administration Potential to connect additional access/preservation copies to the same ID Potential to be integrated into automated and low-cost workflows Non-proprietary solution
5	Which ID scheme and why? Cooperation with a trusted, public and non-profit organization (sustainability – lower risk) Management of a resolution service, other metadata services and an archival copy within the same framework (self-interest – lower risk) Possibility to use the same PID for different manifestations of the same content Sustainability – beyond the http protocol
6	Which ID scheme and why? Primary decisions: Decision to cooperate with the National Library of Sweden (sustainability, self-interest) Decision to fit all needs into an automated workflow (low cost, sustainability) Secondary decisions: Decision to use URN:NBN as a primary persistent identifier Decision to develop a resolution service
7
8
9	The implementation strategy Creating a simple solution that works is better than waiting for a perfect solution It makes it possible to introduce the concept of an actionable identifier to users and to demonstrate its power in a practical way (If component based system development is used ) When new demands, more resources or new technologies appear, some of the components – or the entire application – can be replaced by enhanced ones
10	Resolution Service (RS) application Core functionality Provides resolution by identifying valid URL(s) for resources which have been assigned an identifier within the URN:NBN namespace Provides indirect resolution for other id-schemas which have been registered http://urn.kb.se/resolve?doi=10.1016/S0021-9797(03)00401-6 http://urn.kb.se/resolve?hdl=10.1045/may2001-kahn Additional functionality Provides details about the authority which has been assigned a sub-namespace within the national URN:NBN domain Can store standard metadata for each object together with its assigned ID
11
12	It’s all about the service The quality of the service relys on the fact that relationships between the ID and the associated location(s) of the object (in Internet context, URL) are up-to-date Relevant questions are: How to maintain these relationships (mappings)? How to make updating of mappings between location and identifier low cost and reliable? How to guarantee such a service in the long term? (In Internet based digital library environment PID –> URL, in physical library environment PID –> shelf mark)
13	Example of an implemented solution Integration in a publishing/depositing workflow IDs are assigned automatically Mappings are made available to the resolution service for updating urn:nbn:se:uu:ub:epc-schema:rs-location-mapping Updating of the central service The RS application harvests relevant data to update the mappings & possibly metadata The RS application can merge information about multiple locations referenced by the same ID
14	Implementation Java, XML, XML Schema, JSP, HTML Open Source In operation at NL in Sweden, in Denmark installed at Århus Staatsbibliotek, in Finland installed at the Helsinki Technical University Originally planned to build a network of services based at NL’s in Scandinavia Not implemented (organizations and political level issues)
15	Some problems we have encountered Lack of expertise or staff to maintain the service at national libraries (Finland, Iceland and Sweden at the beginning) Lack of interest for “outside” solutions from the NL (Denmark, Norway) Difficulty cooperating effectively with NL outside of Scandinavia Lack of positive interest and desire to understand from projects running other solutions (defensive position about our choice of URN:NBN) Socially based problems Not really technical ones
16	Future in the context of DiVA Some constraints in the current approach The NBN should be assigned only to resources which the NL is interested in delivering and archiving There is need to use an ID schema which supports all kind of objects Some strengths The DiVA systems supports various identifier schemes through an identifier agnostic element
17	How does this relate to a global resolution service? The methods and tools developed for resolution of NBN supports other ID schemes by indirect resolution (pointing out to other services) by direct resolution under condition the mappings can be harvested by RS The architecture is designed to be distributed ? Scalability ? Willingness of other resolution service providers to exchange data