This paper describes a “generalized Petri network simulator” which can be used to evaluate the correctness and efficiency of communication protocols in distributed systems. Following a statement on the significance of the problems, the authors introduce the Petri network notation and explain the approach. The FORTRAN 77 simulation program is described, an example of a token passing program is given, and an example of a token passing protocol model is presented. The Petri net graph is used to explain the FORTRAN procedures (subroutines), though little information is given on the statistical details of the simulation.
While the subject of this work is of intense current interest, the paper has some major flaws. The defense of the use of FORTRAN 77, a most unseemly choice, provides the most obvious example of the paper’s deficiencies (pp. 401–402):
As is known a discrete simulation technique can be based either on specialized programming languages, as for example SIMULA and GPSS, or on techniques which do not introduce unnecessary determinism, permit possible parallelism and which unambiguously can represent the behavior of the process and of the environment in which this process must be implemented. However, the use of specialized languages firstly needs the availability of the relative compiler on his own computer, then it imposes some further restrictions because these languages are conceived for general purpose aims. On the contrary we intend to propose a simulation approach for computer networking studies written in FORTRAN. . . .
If this defense is both “crystal clear” and technically acceptable, especially the limitation of “general purpose aims [in] specialized languages,” then you may wish to read the paper.