Meta data Driven Website Generation - Tools for Describing Projects
Wolfram Sperber & Robert Roggenbuck, Zuse Institute Berlin
Slides
Abstract
The generation of meta data for webpages is not as simple as it might sound.
Especially when you are focused on a better retrieval quality,
which leads to the need of fine-structured meta data of high quality and the usage
of standardised vocabularies for the content description.
The aim of the project "Math&Industry" (http://www.mathematik-21.de/)
is to generate websites for projects in the area of applied mathematics that
are placed in the difficult border area between mathematics,
the application area and often other needed science fields for the modelling of the problem.
So there was a need to generate webpages, associated with high quality meta data,
for a whole complex website (not only a single page).
Therefore tools were developed to create webpages for describing different contents.
The development of the tool was designed according to two major premises:
1st the tools are made for the use by experts in various science fields, but not (!)
for computerspecialists, and 2nd the requirements of the data input were not determined
by the look of the webpage, but by the needs for a detailed description of the object
and needed information to build further services upon it.
Because of this you can say that the visible webpages are generated out of the stored meta data.
Eight specialised online tools (MIPMs, Math&Industry Presentation Maker),
as well as a content management system were developed to organise the website generation
(WebSiteMaker). The tool for generating person descriptions (MIPMPers),
i.e. professional homepages, and the tool for generating a glossary (MIPMGlossary)
are presented in detail.
The meta data is coded in RDF/XML by using a broad number of standardised meta data vocabularies
(e.g. RDFS, Dublin Core, OWL and MARC).
For the special needs of Math&Industry it was necessary to formulate own vocabularies -
often as refinements of existing terms.
|
