Topic #2 --------------- OP-SF NET ---------------- November 9, 1995 ~~~~~~~~~ From: Charles Dunkl
Subject: Report on SIAM meeting, Charlotte Report on Annual SIAM meeting, Charlotte, North Carolina, October 23-26 Items pertinent to the activity group OP/SF at the business meeting of chairs of Activity Groups with the board of SIAM: - WWW page: SIAM is setting up links to home pages created by the activity groups. Our group intends to start working on a home page very soon (see Topic #1). - 1996 annual meeting in Kansas City, Missouri, July - the theme has been announced - it is "New Tools for Applied Mathematics". This might be a good fit with the topic of the "Handbook" project (see OP-SF Net 2.4, Topic #10) if one considers it as a method of using modern media, such as CD-ROM, to disseminate information about special functions. Furthermore, our group is encouraged to submit suggestions for plenary speakers (45-minute lecture) for the meeting. Such suggestions should be developed by December 15 (and can be sent to Martin Muldoon ). Observations at the minisymposium "Computational Aspects of Special Functions and Orthogonal Polynomials" (see OP-SF Net 2.5, Topic #4 for the program): -John Boyd (University of Michigan) talked about Hermite expansions, the anharmonic oscillator, and perturbation theory. - Walter Gautschi (Purdue U.) discussed an algorithm for generating polynomials orthogonal with respect to a Sobolev norm. This theory has significant differences from the ordinary one, for example, there is no three-term recurrence. - Dan Lozier (Nat'l Inst. of Standards and Technology) gave a survey on software for special functions. He is working on a project to provide accuracy checks for algorithms. The idea would be for researchers to send in a list of pairs (x, f(x)) where f is a special function, and then the NIST computers check these values independently and generate an error report for the researcher. Dan has set up a Web site and asks that the special function community read over his suggestions and send feedback. See Topic #12. Direct link to testing service - Nico Temme (CWI, Amsterdam) talked about the problems of accurate computations of probability distribution functions in regions of rapid increase, for example, incomplete beta integrals with very large parameters. Often such computations require much more work than routine problems. - The writer of this note presented joint work with Don Ramirez on computation of surface measures of ellipsoids in N-space via Lauricella F_D functions and an application to optimal designs in statistics.