Third Workshop on

Very Large Digital Libraries

September 10, 2010

In conjunction with the 14th European Conference on Research and Advanced Technologies on Digital Libraries (ECDL)

What is a Very Large Digital Library (VLDL)?

In fact, there is no clear and well-accepted definition. Some, in an attempt to give one, blur the separation between Very Large Databases (VLDBs) and VLDLs and consider the latter as VLDBs storing Digital Library content – VLDLs are DLs storing content beyond a given threshold. Nevertheless, DL systems practice seems to prove the limits of this intuition, whose focus is only on “very large size of content”. Indeed, as motivated by the DELOS reference model for DLs[1] , user management, functionality management, and policies are aspects as important as content management in the definition of a DL system. In fact, VLDLs should be better described as DLs featuring “very large” issues in one or more of these aspects. Examples are: dynamic integration of arbitrary numbers of DL functionalities, very large numbers of user recommendations to be processed, or very large numbers of DLs to be federated under the same integrated policies. Moreover, being dependent on more aspects, the measure of “very large” for DLs seems to be more variegated than for DBs (nowadays a database is very large if its content is beyond 1TB of data). For instance, interoperability issues arising when integrating and federating different DLs, introduce new forms and measures of “very largeness”. As an example, consider VLDLs dealing with the dynamic aggregation of an arbitrary number of possibly heterogeneous OAI-PMH compatible repositories. Such systems are not concerned with the byte-size of the aggregated content, but rather with the very large numbers, e.g. hundreds, of repositories to be integrated in an systematic, automatic, effective, dynamic, possibly reversible way. Finally, addressing sustainability of VLDLs also introduces interesting research problems. Indeed, VLDLs are supported by organizations, e.g. libraries, whose resources are often budget-limited and cannot include, for example, skilled programmers, powerful hardware or high-speed connections. This scenario calls for research on advanced software systems capable of meeting VLDL requirements under low-cost constraints.

In summary, research on VLDLs opens up novel and actual scenarios, where researchers have to confront with new foundational and system design challenges in a context having scalability, interoperability and sustainability as focal points.

References

For any information you may need, please contact the .