.AM 19.num/RJF.mitchell.etal .ls 2 .na .LP Topics in the Representation and Manipulation of Mathematical Knowledge Nicholas Mitchell, Dr. J. Daniel Halpern*, Ben Berman, and Chien-Long Lin (Professor R. J. Fateman) (NSF) CCR-92-14963 We are investigating computer technology in support of the representation and manipulation of mathematical notation by computer. Our long-term goal is to produce a consistent and constructive representation of a substantial, useful class of engineering and applied mathematics, driven by the needs of significant applications. Key issues we hope to examine include the use of multiple representations, and the encoding of geometric, algebraic, logical, algorithmic, and physical relations in symbolic mathematical computation. We are working toward enhancing tools for symbolic manipulation interactive systems to increase the productivity of users of .I pencil and paper .R mathematics\(emincluding physical scientists, programmers, and researchers. Some specific recent topics include: Development of new algorithms and representations for manipulation of functions of a complex variable correctly and with some generality; in particular, both correct manipulations and proofs in analysis must account for branch cuts. Provision, through interval-arithmetic, of "honest" plots of mathematical functions guaranteed to be free of the defects present in usual plotting programs that are defeated by singularities, periodicities that match the sampling frequency of the plotting program, and so forth. Representation and use of large-scale tables of indefinite and definite integration formulas to supplement algorithms. Experimentation in optical character recognition and two-dimensional equation parsing for the automatic input of reference material. Improved pattern matching and search in discrimination trees for algebraic patterns. Improved numerical code generation from a traditional symbolic computing language (Lisp) to provide for more natural treatment of symbolic-numeric algorithms. *Post-Doctoral Researcher