Gerhard-Mercator-Universität  Duisburg  Department of Mathematics

Within the first year an infrastructure was set up and a number of problems were solved, much renained to be done. The second year started with a heavy emphasis an the colaboration wth libraries and unibersity computing centers. Our project is structured into several parts that form together an integral approach:
 
 

Metadata	Retrieval	Formats
Multimedia	Support	Libraries (DDB and SUB)
	Coordination

Since dissertations assume at least a double function, i.e. on the one hand represent current scientific publications, and on the other hand they are probably one of the most important documents during an individual qualification process, metadata, i.e. structure information about the appropriate work, have to meet defined requirements.

Thereby one has to distinguish between bibliographic and department-specific metadata on the one hand (library package), as well as examination-relevant metadata on the other hand (rights package). Regarding provision of data one has to distinguish between data given by the author  (e.g. titles and abstract of the work or also personal data of the author), and data given by the library. Furthermore, suitable metadata must be collected, which, among other things, authorizes the thesis version in order to ensure the security of the work regarding copyright. In the context of this project suitable metadata packages are now being produced on the basis of the Dublin Core and are to be coordinated within the special interest communities.
Our HTML4-coded metadataset is authorized by the German National Library (DDB) and every graduate student is obliged to give the dissertation specific metadata to the DDB if he wants to publish the dissertation electronically.
Within our subproject we developped an entry form for digital dissertation metadata, which is used in the participating projects. There is a complete documentation with the DDB and a well structured table of dissertation specific and technical metadata. In accordance to our metadata set we developed a broker interface based on Harvest.

These results may be quite remarkable, but it is not yet sufficient for a reasonable procedure. For example it is not possible to create a metadata file for a dissertation made by a group of graduate students in the HTML4-Code because no one can assign the names to the e-mail addresses in an unambiguous way. And that is only a little problem amongst other greater problems yet to be solved.

Because of these problems we started to make a proposal of a RDF-metadata file.
As you can see there, we need a great part of all DC-Elements. But there are still some problems.
We need for example some Titletypes like "Main" for the tile of a dissertation in contrast to the subtitle or alternativ title.
Because of the political aspects, dissertation specific metadata had to be defined in an unambigious way.
 

• Qualifiers:

Titletype:	Main and Alterative
Creatortype:	PersonalName, PersonalName.Address, PersonalName.DateOfBirth and PersonalName.PlaceOfBirth
Publishertype:	CorporateName and CorporateName.Address
Subjecttype:	Topic and Keywords
Descriptiontype:	Notes
Contributortype:	Advisor, Referee, CorporateName and CorporateName.Address
Datetype:	Submitted, Created and Accepted
Relationstype:	IsPartOf and Contains

 

 
• Schemes

 

Subjectscheme:	Thesaurie
Datescheme:	WNT8601

 

---