Thesis Work (1995 - 2008)

The systems I built for my three theses, from the PhD work back to the project thesis. DLMAPS, VISCO and GenEd have their own pages with screenshots, publications and more detail. The titles link to the source code on Github.

2001 - 2006 (PhD Work)
DLMAPS (Description Logic Maps) - source code on Github, and the project page on this site.
An experimental, prototypical description logic-based geographic information system, featuring a high-performance deductive query answering engine. The framework supports hybrid (spatial and logical) representations and is capable of processing ontology-based spatio-thematic queries.
 
The Racer query language nRQL is an instantiation of this generic query engine framework.
 
Similar systems have appeared, e.g. "Pellet Spatial" ™ by Clark & Parsia, and standards such as GeoSPARQL are emerging.
Main papers: Journal of Applied Logic, ESCoR 2006, KRDB 2003. The full account is in my PhD thesis.
2003 - 2005 (PhD Work)
MiDeLoRa (Michael's Description Logic Reasoner Framework) - source code on Github.
A framework and domain-specific language - a set of Common Lisp macros - for constructing description logic reasoners. It comes with tableau provers for a number of standard description logics, and works directly on the data model of the DLMAPS substrates.
Described in the Journal of Applied Logic article and in chapter 6 of my PhD thesis (see the English summary).
1997 - 1998 (Diploma Thesis)
VISCO - source code on Github, and the project page on this site.
An experimental visual spatial query language and prototype system for geographic information systems, applied to digital city maps. The VISCO spatial query processing engine featured a cost-based heuristic spatial optimizer, similar to the nRQL engine later (see pages 165 - 180). Perhaps you will enjoy our demo video ;-)
Main papers: VL 1997, VL 1998, VL 1999.
1995 - 1996 (Project Work)
GenEd - source code on Github, and the project page on this site.
A description logic-based graphical editor for designing visual formalisms (e.g. Petri Nets). The editor is capable of recognizing spatial constellations described by spatial concepts in a description logic ontology. Enjoy our demo video ;-)
Main paper: VL 1996. Thesis: project thesis (in German).