Glenn BrunsEnabling Computing Technologies
In applying computer science, one general approach I like is abstraction, in which analysis is made tractable by ignoring details not relevant to the property of interest. My PhD work concerned abstraction techniques for verifying temporal properties of processes, and my work (with Patrice Godefroid) on multi-valued model checking is also in this vein. I would like to do more in the area of designer-guided abstraction, where the designer of a system can provide hints or annotations to show which parts of the system design are relevant to key system properties.
I also like constructive approaches to achieving correctness and other system properties. The idea here is that one should achieve desired properties of a system through the construction process, not through (an often intractable) post-hoc analysis. This is one of the reasons I like domain-specific languages and code-generation.
Warning: the photo above is ancient (1998)!