December 1966 - Vol. 9 No. 12

This report on the author's trip to universities in Western Europe in the summer of 1966 gives brief descriptions of computing activities at each institution visited. Present equipment capabilities vary from moderate to large scale; however, many institutions plan to acquire complex time-shared systems in the near future. In the author's opinion, the state of the art lags behind that on this continent. This lag is attributed to four principal factors: (a) the handicapping organization of academic procedures; (b) the university-government financial relationship; (c) the subordinated organization of the computing facility; (d) the paucity of professional interchange of knowledge. The effects of these constraints are explicated.

The FLOWTRACE system produces flowcharts of programs written in “almost any” programming language. One must describe the syntax of the control statements in his language; for this purpose a metalanguage is available. The resultant object deck is used to flowchart any programs in the language described. Several examples of FAP and SNOBOL flowcharts are given. However, it is not necessary to confine one's scope to existing languages. One may define his own language in any “well-structured” manner. This feature is particularly useful when it is desirable to chart only comments within a program. Such an approach permits the documentation of descriptive remarks and avoids the inclusion of coding details.

The evaluation, by the propagation of variance technique, of the sensitivity of time-varying systems to initial condition and parameter errors, involves the determination of several system-dependent partial derivative matrices. This requirement has led to separate programs for each system under investigation. A new program, through utilization of the Wengert differentiation technique, automatically determines the required matrices from specific system equations supplied in subroutine from at execution time, eliminating the need for individualized programs, and presaging the further development of extremely general computer programs.

The second Remes algorithm as originally established for polynomials, may converge or not when the approximating functions are rational. However, the few results known in this domain show how efficient the algorithm can be to obtain approximations with a small error, much more than in the polynomial case, in which the best approximation can be very nearly approached directly by a series development. The aim of this paper is to investigate the limitations of the applicability of certain extensions of the algorithm to the case where the approximations are rational as well as to present some numerical results.