This Master's Thesis project should focus on making existing (rather complex) software accessible to working dialectologists. It thus emphasizes usability and the user interface.
Alfa-informatica researchers have developed new techniques for measuring the differences in pronunciation. The novelty in the techniques is the fact that they produce genuine measures, i.e., numbers, which correspond to the degree to which pronunciations differ. Wilbert Heeringa maintains a dialectology site which illustrates the main ideas, using Dutch dialects as an example. His site contains pointers to publications about the work and also to similar work by others.
As part of a grant from the Dutch National Science foundation, Peter Kleiweg has created a powerful UNIX package -- implemented in ANSI C, and making extensive use of an interface to the statistics package R -- so that others can experiment with these ideas. The Levenshtein Software Site is illustrated with material from American dialects. See Kleiweg's further notes for an idea of how the system is to work.
The problem is that dialectologists don't use UNIX, and don't, in general, know how to program. The task of this project is to select a usable subset of the UNIX facilities and to create a Windows-based package which will allow dialectologists to perform at least some analyses, preferably as many as possible. The boundary conditions are auspicious: Mr. Kleiweg is committed to providing platform-independent C programs which should compile in Windows directly, and two local dialectologists would like very much to serve as guinea-pigs (beta-testers) for a Windows-based package.
The project will be successful if a good deal of functionality can be provided in a way that the non-programmer users can exploit.
This project would be most suitable for two students working together.