Mathematics and Computers 0 North-Holland Publishing
AUTHOR’S
in Simulation Company
XXII (1980)
10
REPLY TO COMMENTS BY POSCH AND SCHMIDT
D.M. AUSLANDER Department
of Mechanical Engineering,
University of California, Berkeley,
CA, U.S.A.
In item 1 (ii) it is claimed that “. . . sorting algorithms in simulation packages seem to be of little use . . .“. To the contrary, with respect to the kind of structure used by PARASOL, the sorting algorithms perform a useful function and one that is consistent with the point made by Posch and Schmidt that, “The syntax of the language used should permit the formulation of mathematical expressions in their original form . . .“. In ACSL, for example, a commercially successful CSSL, the sorting ability is a central feature of the language. Finally, a comment on the way in which programming time was determined for the ‘none’ listed as the programming time in PARASOL for PHYSBE. ‘Programming time’ here is meant to refer to the time elapsed between completion of the user’s analysis and understanding of the mathematical model to the time he/she is ready to enter the program into the computer system. Understanding the mathematical model is not included in programming time because it is independent of the simulation language to be used; the time required to enter the program into the computer tends to be more a function of the nature of the computer that the language is implemented on rather than the language itself, so that is not included in programming time either. In the case of PARASOL/PHYSBE, no intermediate work is required between understanding the mathematical model and entering it into the computer. This assumes a user already familiar with PARASOL.
I would like to thank Drs. Posch and Schmidt for their comments, indeed, their extended discussion, on simulation language my paper, “A continuous-system for LSI economics”. In the spirit of variety presented in the original paper, I find myself in agreement with most of their comments. That is, no one form of simulation language (or other languages, for that matter) can meet all needs and all users’ personal preferences. By making strong use of manufacturer-supplied software, languages like PARASOL or INTERSIM can be econimically developed for primary use within a single user community and designed to meet the needs of that community. Within this context, I woulk like to make some ‘comments-on-comments’, concerning some of the points raised by Posch and Schmidt. In particular, I think that they overstate the difficulties involved in making use of the language structures used in PARASOL. I can appreciate and respect their choice of structure for INTERSIM. Like INTERSIM, however, PARASOL has been in successful use in an educational and research environment for several years, solving a variety of purely simulation and real-time problems. With respect to the RPN format, for example, I have found that users have little or no difficulty learning or adjusting to it. ‘Compiling’, to use their word, algebraic notation to RPN can be done strictly mentally, from equations handwritten on paper to typed-in PARASOL code. RPN to algebraic is not quite so obvious but, also, causes little difficulty.
10