A methodology based on recording a logical flow of steps associated with library functions is described in this paper. In this way, a main program can be made independent from step content by basing logical flow on parameters between the transfer of steps. The paper explores improving program flexibility by storing and reusing functions without touching the source in the high-level language.
I had a number of issues with the paper. First of all, the verbose, hard-to-understand, and occasionally incorrect language was sometimes challenging. Second, the examples are based on antiquated flow charts. It would be usefulto relate this to the closest equivalent modeling technique in unified modeling language (UML) activity diagrams. Third, the comparison to component and rapid prototyping methodologies could have included dataflow languages, such as PGM or Autocode, or dataflow tools such as Sage, Lockheed Martin’s Graphical Entry Distributed Application Environment (GEDAE), and those by Management Communications and Control, Inc. (MCCI), since that is the context the analysis, design, and programming technique is based on. A comparison of the project’s future aspirations, of connecting the logical icons of dataflow with the aforementioned time-proven dataflow techniques, would make the paper’s contributions clearer.
A multiple pass technique is described: the memorization of sequence steps that are later managed by a main program. Other multiple pass methodologies, such as profile-guided optimization, improve performance in an application-specific and lower-level fashion. The main contribution of this paper is the description of a novel way of lifting logical dataflow steps to perform a more abstract form of sequencing.