APPENDIX 2
Proposed Use of an NLA Re-direct Service

1. Requirements of a Persistent Identifier scheme

The requirements for a long-term fully operable persistent identifier scheme are as follows:

The resolution process needs the following capabilities.

In addition, identifiers for materials archived for long periods (created by one organisation and copied for preservation by another) must be “transferred” or routed to the new organisation for resolution to the archival copy when a publisher has given up responsibility for the work. There must also be a reverse look-up capability.

Most of the PI systems developed for use on the Internet currently fail one or more of the criteria above and there is no contender sufficiently ahead of the others to rule out of the others as candidates for implementation. Various institutions are taking a variety of measures to deal with the issue of persistence.

The most significant failure of the systems lies in the area of browser recognition and routing. The identifier has to be recognised, a server registered as capable of resolving the identifier has to be located and the identifier routed to that server in order for the server to return the resource or metadata. The first two steps require a high degree of interoperability and the widespread deployment of an infrastructure based on open standards. The method of resolution employed once a request reaches the host server is generally considered to be a local issue and is transparent to the requester.

Most systems which incorporate a resolution capability use a simple database to associate the Identifier with the actual location of a resource when a resolution request is made, e.g. PURL, Handles, DOI.

A proposal for a re-direct using NLA applications

The digital surrogates at the NLA are delivered by applications that are database driven. These applications have indexes in which persistent identifiers are linked to the locations of the resources. An application must always “know” where the files or digital objects are for the delivery of the service. Because they are the basis of service delivery applications, the databases will always be updated when resources move or applications change. They are therefore also capable of doubling as internal resolver databases for servicing requests based on persistent identifiers.

Tony Boston has developed a proposal that this capability be exploited to resolve URL style identifiers for NLA resources. The proposal involves the development of a general server side program that redirects queries for a particular digital object to the appropriate application based on the type of object being delivered, e.g. if the item were a picture, the request would be directed to the Images1 database, if a manuscript, to whatever delivery mechanism is developed for manuscripts. The program would do this using information in the URL that would indicate the type of material the identifier was describing.

Advantages of the proposal

Low risk

The proposal is a low risk option for digital surrogates. It involves the development of a (relatively) simple program to parse the identifier so that a request can be directed in the form of a query string to the appropriate application. It is essentially a persistent URL using a different form of redirect.

Easy maintenance

The proposal would be easy to maintain. The applications are active delivery mechanisms for the digital resources and therefore databases must be updated when there is any change of application or change of location of the files. Minor changes only would be needed to the redirect programs if there is a change of application software.

Compatibility with future options

A complete unique identifier for the resources will be stored in the URL so that it is not dependent on the domain name of the URL to provide uniqueness. This ensures that it does not rule out later use in the implementation of another system such as DOI, Handles, ARK etc. Indeed, with a slight modification, it could be used to populate a database associated with one of the other options.

Speed of implementation

The proposal would be quick to develop for immediate applications for the materials being produced for the digitisation program.

No new database required

It does not require the creation of a new resolver database as would be necessary for the implementation of Handles or DOI or OCLC’s PURL.

Particularly suitable for digital surrogates

The proposal is particularly useful for digital surrogates for the following reasons:

Disadvantages of the proposal

The system is a persistent URL and is therefore subject to the same disadvantages as a PURL. It is not protocol and location independent. The method could be implemented in a similar fashion to a PURL.

The URL is re-directed, so the URL displayed in the browser is the destination URL (a query string) rather than the source URL. This is a problem with all resolving services that are based on a re-direct, eg PURLs, Handles or DOI using a proxy server etc.

It does not work well for current, dynamic websites that are not database driven.

It is of doubtful value for Pandora or archives of resources which do not originate with the archival institution because it does not have a multiple resolution capability. It would be difficult to support one to many resolution, e.g. multiple instances or multiple versions.

It does not offer a basis on which to offer a national persistent identifier system. It is a single site solution dependent on the existence of a database application to access the resource. (Although it could be cited as an example of a resolver database and the NLA could offer to assist those who wished to follow the example.) It should be noted that this method is being proposed as a method of managed URL or persistent URL for NLA resources and is not being proposed as a universal solution to the general issue of persistent identification.

Comparison with other persistent identifier schemes considered

The following table is a rough comparison of the option with other persistent identifier schemes.

PID Location Independent Protocol Independent Browser recognition Routing Relative cost Relative risk Multiple resolution
Proposal No No Yes http/DNS Low Low No
PURL No No Yes http/DNS Low Low No
Handles Potentially Potentially No. Plug in Handle servers Medium Medium Potentially
DOI Potentially Potentially No. Plug in Handle servers High Medium Potentially
URN Yes Yes No No ? ? Potentially

Recommendation

That a general server side program that redirects queries for particular digital objects to be resolved by the appropriate application software be developed for the NLA’s digital surrogates. The program will achieve this by parsing a URL style identifier.

[NB. This proposal was accepted by the Persistent identifier group, April 2001]

Return to Table of Contents | Proceed to Appendix 3
About This Site | Copyright | Privacy | Accessibility | Site Map | Site Search | Content A-Z | Contact Us