Article review: A message-based approach to discrete-event simulation by BAGRODIA, RAJIVE L. (Univ. of California, 8807-0549 Los Angeles); CHA~DY, K. M.; AND MISRA, JAYADEV (Univ. of Texas, Austin). IEEE Trans. Softw. Eng. SE-13, 6 (June 1987), 654-665

Abstract

I I] As a service to our readers, SIGSIM has reached an agreement to reprint reviews of books and papers on simulation and modeling that originally appeared in ACM Computing Reviews. CR is a monthly journal that publishes critical reviews on a broad range of computing subjects including simulation and modeling. As an ACM mem~r, you can subscribe to CR by writing to ACM Headquarters. A message-based approach to discrete-event simulatioe. The authors clearly state their goals as (I) developing a message-based approach to discrete-event simulation, (2) showing that the addition of a small language fragment permits general purpose languages (GPLs) such as FOR-TR.A.N and Pascal to join a "'class of simulation languages ," and (3) giving an example of such a language, MAY, derived from FORTRAN. With the intent of accomplishing these goals, the authors explain the concept of message-based simulation and their design philosophy, note the constraints needed using some examples, and describe an implementation in MAY. This paper is disappointing. Kristen Nygaard (co-developer of SIMULA with O. J. Dahl) once remarked to me that "only the people who used SIMULA for simulation applications really understood the power of the language." In a larger context his criticism was directed more broadly at the softy, arc enginering community that rushed to embrace SIMULA as the first language implementing the abstract data type concept while ignoring equally important characteristics and capabilities. The software engineering community continues to ignore the simulation programming language (SPL) and model representation developments; otherwise, this paper would not have been published. This paper contains (I) conceptual misunderstandingsmSIMULA does provide message-passing but not in the explicit form preferred by the authors; (2) incorrect statements~a SIMULA object does not cease to exist when no references to it remain if it has not exhausted its operations (s~ [I], pp. 48-49); and (3) undefined terminologym'a small programming language fragment" is the crux of the transformation of GPLs to special class SPLs, but a general definition of a fragment is never given. Yet the most serious deficiencies of the paper lie in the ignorance of key issues that supposedly motivate the message based approach: (1) The authors' entity offers nothing beyond the SIMULA process. The generalization of entity to encompass both resources and processes is borrowed directly from the SIMSCRIPT II permanent and temporary entities. (Note that even the terminology is the same.) (2) The wait-umil (Boolean) construct has been a …