European Library Automation Group
ELAG 2007
Barcelona (Spain), 9-11 May 2007
http://elag2007.upf.edu/
 
  
 

A prototype of a Digital Library Management System
Heiko Schuldt (University of Basel, Switzerland)
Vittore Casarosa (ISTI-CNR, Pisa, Italy)

  

In parallel with the efforts for defining a Reference Model for Digital Libraries, DELOS, the Network of Excellence on Digital Libraries, has started a global integration effort for the implementation of a prototype of a Digital Library Management System, which can show combined text and audio-visual search functionality, personalized browsing using new information visualization and relevance feedback tools, annotation and processing of retrieved information, integration and processing of sensor data stream. From a systems engineering point of view, the DELOS DLMS allows simple configuration and adaptation, while being reliable and scalable. It can be considered as a first step towards the implementation of a system embodying the ideas and principles being defined by the DELOS Reference Model.
  

The prototype is being built by integrating digital library functionality provided by the DELOS partners into the OSIRIR/ISIS platform, i.e., a middleware originally developed at ETH, and presently being extended at the University of Basel. The OSIRIS middleware (Open Service Infrastructure for Reliable and Integrated process Support) supports programming-in-the-large; i.e., the combination of arbitrary application services into so-called processes. This is realized by a set of generic (application-independent) services that include the registration of services and processes, interfaces for application development, an engine for decentralized execution of processes, and services for load balancing. In addition, it features reliable execution by applying advanced database concepts at the level of processes for failure handling and concurrency control.
  

ISIS (Interactive SImilarity Search) is a set of Digital Library services that have been developed on the basis of the OSIRIS middleware. ISIS includes a sophisticated index structure (VA-file) for similarity searching, which is particularly well suited for high-dimensional vector spaces. Furthermore, in terms of Digital Library functions, ISIS features support for content-based audiovisual searching and rudimentary textual searching. It also provides basic support for relevance feedback and visualization.
  

The services that have been integrated or are being integrated into the DELOS DLMS are the following.

• Interactive SOM-Based Retrieval (use of the Self Organizing Maps to help users in browsing and retrieving);
• 3D Content Based Retrieval (retrieval of 3D objects based on similarity);
• Audio Retrieval (retrieval of music based on similarity);
• Advanced user interfaces (the Daffodil system and the MedioVis system; visualization of structured meta data with the DARE system);
• Active Paper (interaction with the digital library through a digital pen and special paper);
• Annotation Service (annotation of digital objects and search for annotations);
• Natural Language and Speech interface;
• Semantic Video (semantic storage and retrieval of video content);
• Video Annotation;
• Video Retrieval by content;
• Indexing Term Extraction (sophisticated term extraction from multi lingual text documents);
• XML Query (improved search functionality for XML content);
• Peer to Peer Search Engine (retrieval of text and images from a P2P distributed environment);
• Alternative user interface (for the naïve user).

For testing purposes, the DELOS DLMS includes a number of image and video collections, assembled from various sources. The main ones are the following:

• ISIS collection, about 54'000 images with mixed contents (cars, space missions, paintings, drawings, commercial photo stocks);
• ETH World collection, about 625'000 images and text files from the websites of ETH;
• Medical images collection, about 50'000 images with medical contents (radiology, nuclear medicine, pathology);
• Art gallery, about 16'000 images from the Web Gallery of Art;
• 3D collection, about 1'000 3D models of different objects;
• Audio collection, about 1'000 audio files (speech and music);
• Video collection, about 1'200 video sequences, mostly with contents related to sports.

The main purpose of this workshop is to stimulate discussion on the overall functionality and architecture of the "ideal" Digital Library Management System. The outcome of the workshop will be useful input for planning further development of the prototype and, of equal importance; it will provide valuable feedback for the Reference Model and the Reference Architecture.
   

In each of the first three "sessions" of the workshop it is planned to briefly introduce and demonstrate the topics to be discussed, starting with the architecture and the functionality provided by the supporting middleware (the OSIRIS system) and continuing with the tools and services presently available, in order to start a discussion on the relevance, the advantages and the disadvantages of each component. The final session will try to draw some conclusions about the architecture and functionality of the system, providing a prioritised list of issues and recommendations.