00100	NOTES ON RESEARCH TOPICS, ESPECIALLY FOR THESES
00200	
00300	
00400		I am prepared to supervise research in the following topics,
00500	although I am also prepared to consider other topics proposed by
00600	students.  Which of these topics would lead to PhD theses is
00700	not immediately apparent.  In most of them, the student would have
00800	to contribute substantially to the definition of the problem as
00900	well as to finding a solution.
01000	
01100	SYSTEM PROGRAMMING
01200	
01300		It would be hard to persuade me to let a student design
01400	a new programming language or design an operating system as
01500	a thesis project.  There would really have to be a new idea
01600	and not merely a dissatisfaction with the details of present
01700	implementations.  Instead I would prefere to see research in
01800	new areas such as the following:
01900	
02000		1. Information description.  Many files exist that
02100	are accessible to our computer over the ARPAnet or by
02200	telephone.  Develop a language for describing a class of
02300	such files that would permit a program to answer a question
02400	by extracting the information from such a file.  Easy (I hope)
02500	example:  many ARPAnet sites have telephone directories in
02600	their computer as do we.  These directories all have approximately
02700	the same information but have different file formats.
02800	Devise a way of describing such files so that a single program
02900	can find addresses and telephone numbers from any of the files
03000	using the description of that file.
03100	
03200	#. Business communication language.  See CBCL[f75,jmc].
03300	
03400	THEORY OF ARTIFICIAL INTELLIGONCE
03500	
03600	#. Minimal models of a collection of sentences
03700		See me for an oral spiel, but first read minima[f75,jmc].
03800	
03900	#. Finitization.  A possible scheme for showing certain sentences
04000	satisfiable by translating them into theories in which satisfiability
04100	means satisfiability in models of a fixed size.  See finiti[f75,jmc].