As a research associate at the Chair for Information Technologies it was my primary task to design, implement, launch and supervise a companywide knowledge management (KM) project at the Fraunhofer Institute for Integrated Circuits (IIS).

In the project, I decided to focus on knowledge identification (Who knows what?), because it seemed among all the KM-stuff going on at that time most promising to become any (minimum) kind of success story.

The project was and still is subject of scientific reseach, too, and led to a number of publications.

As part of the project, I first of all designed, programmed, launched and managed a new, webbased, companywide intranet.

From the technological point of view, I basically combined the web application system ZOPE with its great "through the web programming" capability, MySQL as backend database for relational data and mainly Python as programming language.
Thus I got a programming environment, which made extremely fast software development, testing and deployment as well as reaction to user demands possible.

Now, in order to make the employees use the intranet at all (me and from 2006 on a collegue) we implemented numerous, usually very simple (but effective), webbased applications (e.g. room reservation service, news service, webbased document transfer service, multiple ticket systems, content management software, ...).

Although, at first glance, these applications have nothing to do with KM, they made the employees use the intranet regularly.
This way I could use the intranet as marketer and multiplier for the KM-activities and raise awareness about KM among the staff of the whole Fraunhofer IIS.

For KM with focus on knowledge identification, first of all I built a topic map in order to standardize the meta-description language for knowledge and documents. Topics serve as index terms for both. By now, the map has more than 1300 topics.

Then, I put up an expert directory (sometimes called yellow page service, too) and connected it with already widely accepted and used intranet applications.
For the expert directory, employees describe their knowledge by picking topics from the topic map. By doing so, they automatically appear in the directory as possible contact person.

Addionally, I built a document storage and deployment software inside the intranet in order to ease the exchange of "knowledge-intense" documents among the employees.

From the research point of view, I was interested in making statements about the precision of infering from document authorship and document usage (downloads) to knowledge of employees.

Formulated as question: Is an employee knowledgeable on a certain topic if he/she a) writes, b) uploads or/and c) downloads a certain amount of documents related to this topic?

The answer to this question is, of course, very interesting for the optimization of expert finder/recommender software.

With the directory of experts on one hand and the document storage solution on the other hand - both utilizing the same map of topics - I had a research setup which provided real world field data for deep scientific analysis.

Furthermore, while pursuing the KM-project, developing the intranet and a setup for the exploration of the questions mentioned above, I went into fields like data privacy, game theory and motivation theory, which made the whole work an even more fascinating journey.

For more information, please take a look at the list of publications.