I have always had two main branches to my research. The major component is in knowledge representation, and the minor in computer music. There is now a third branch in digital libraries.
The brand of knowledge representation that I favor is conceptual graphs, first developed by John Sowa in the 80s, when he was at IBM. His book "Conceptual Structures", published in 1983, remains a model for a thorough, scholarly account of the issues in knowledg representation, and introduced the AI world to an updated version of the semantic network, that he called a concptual graph. Many people have contributed to its development, and an annual international conference of adherents continues to flourish. I and my colleagues have extended the original concepts in conceptual graphs to cover sets, procedural knowledge, and abductive reasoning. We have built several versions of a conceptual graph system, and Heather Pfeiffer is currently working on its latest manifestation for her PhD, incorporating efficient data structures and parallel algorithms.
See conceptual programming and model generative reasoning for selected publications.
I was involved in computer music before synthesisers and computer-controlled music generation became the norm. In fact, I built my own computer-controlled synthesizer in 1976, complete with standard notation sequencer written in Fortran! It has remained an academic hobby ever since. I have made many attempts to initiate a real research effort, but even though I have had help from several excellent graduate students over the years, it has always remained a low-level affair of the heart.
Since the advent of MIDI, I have taken an interest in its development, and have written several utilties for handling MIDI data in a variety of forms. The main product of this is a visual language for manipulating MIDI files called MidiTrans. Now in its fourth version, for Windows 95 or NT, it has come to be a good test bed for ideas in languages, and music processing, and is capable of handling a wide range of MIDI tasks. You can see some screen shots and a presentation for more detail, or contact me directly for availability of the software.
Sponsored by Los Alamos National Labs, this work involves making archived material of many different sources available on-line to the user community. We will be looking at a wide rangr of issues, from OCR to intelligent tutoring. See the digital library page for more detail.