D Greenstein (DLF) and Martin Halbert (Emory University)
April 18, 2002
Even at their different stages of development, digital libraries are beginning to look for generalizable technical solutions and for an architecture or framework within which such solutions can be implemented, shared, and discussed.
Some of the DLF member library programs (Harvard, NYU, Columbia perhaps) focus from the outset on the design an implementation of such a framework. Others (Michigan, Indiana, Virginia, perhaps) develop this orientation after a period of experimentation that produces a legacy of different, potentially incompatible collections and services.
What is interesting is how so many are now actively engaged in trying to define and build some over-arching digital library architecture. Indeed, it is arguably the one activity that DLF members (digital libraries perhaps) share most in common.
In light of this, DLF is initiating a series of discussions about key architectural and technical issues confronting digital library developers. A proposal prepared by Martin Halbert (Emory University) is supplied below as a contribution to seed discussion at an initial meeting. It focuses on the need to surface a high-level architecture. It is recognized, however, that OCKHAM may evolve in different as a venue where developers share information and insights about a variety of technical approaches and issues that confront them.
The initial meeting will be held on May 10th in conjunction with the DLF Forum in Chicago.
Martin Halbert's contribution follows.
There is a rapidly emerging arena of knowledge management systems that squarely lands in the middle of many of the needs of digital library endeavors these days. Many research libraries are currently having to cobble together our own ad hoc approaches to such systems, often built on top of software that is not based on the emerging standards of the library field. A group of digital librarians believe that a dialogue concering this issue should be initiated immediately, and that our respective efforts could be mobilized around the development of an open framework based on key standards such as METS, OAI, Dublin Core, EAD, TEI, and OpenURL. Systems developed using this framework would embody the modular design of new systems such as the Open Digital Library architecture from Virginia Tech, the National Science Digital Library, and other such systems. Components developed in this framework would share interoperabilty and could interchangeably provide the functionality of portal, metadata repository/PURL server, and content management systems.
Many research libraries have realized that we are individually working on pieces of such a system, and many feel that we are at a point where a face-to-face meeting would give us the opportunity to develop a set of development specifications and API protocols. Our working name for such a framework is OCKHAM, for Open Community Knowledge Hypermedia Administration and Metadata. The intellectual linkage to the philosopher William of Ockham is intentional, as his notion of not proliferating entities unnecessarily is similar to our concepts of not proliferating standards and development efforts unnecessarily if we can collaborate to achieve common goals.
As librarians who are part of a community that has devoted great effort to developing standards, we feel we can surely collaborate on the development of a system that is at least as good as other ad hoc non-standard open source efforts that are being adopted in many arenas, such as the rapidly spreading *Nuke family of content management systems.
And perhaps we can lead the way with a system that walks the walk we have been talking about for years now.