Every year more than one billion gigabyte of information is generated and just 0.003% is printed. Most of these information are pictures, animations, films, sounds, graphics and other multimedia data. Since those data exist only in the electronic form, there is a new challenge for libraries. Today`s digital libraries, however, do not support these data optimally because they are based on the assumption that all documents can be described textually. This causes the risk, that libraries aren`t able to handle multimedia data like graphics, music and e-learning materials sufficiently which results in loss of information for teaching and research.

 

For this reason OFFIS develops a central service center together with the experts from the TU Graz and University Bonn, the Technical Information Library Hannover and the Bavarian State Library. The German Research Foundation (DFG) funds the project PROBADO under the topic "Digitale Text- und Datenzentren zur Sammlung, Sicherung und Bereitstellung von digitalen Quellenbeständen und Datengrundlagen für Forschung und Lehre".

 

The project partners build within a development phase of five years tools, systems and processes, that facilitate libraries to deal with non-textual documents as naturally as they deal with books and magazines today. In the first phase the application domains are in the area of music, 3D-graphics and multimedia teaching materials. Other document types such as pictures and films will be added later.

 

A particular interest in PROBADO is the content-based analysis of non-textual documents. In this librarian workflow procedure a PROBADO-Service will semi-automatically annotate documents with metadata. On top of this documents that are stored in the PROBADO-System get a unique identifier a so called Digital Object Identifier (DOI). These citeable documents bring a huge benefit for authors of non-textual documents.

 

Special innovative user interfaces for the PROBADO-service are in the development process. To formulate queries as well as for presenting the results, new document specific interfaces and processes will be created. The user can e.g. query the dataset of 3D-architecture by scaling the sketch of a house. Beating, humming or whistling are possible queries for investigating the music dataset as well as playing a part of a song on a virtual piano. Next step is to integrate perfected approaches for querying. Enabling the user to browse over all kind of documents is an eminent excess value of PROBADO compared to today`s librarian services.