Structured Analysis (SA): A Language for Communicating Ideas

Abstract

Structured analysis (SA) combines blueprint-like graphic language with the nouns and verbs of any other language to provide a hierarchic, top-down, gradual exposition of detail in the form of an SA model. The things and happenings of a subject are expressed in a data decomposition and an activity decomposition, both of which employ the same graphic building block, the SA box, to represent a part of a whole. SA arrows, representing input, output, control, and mechanism , express the relation of each part to the whole. The paper describes the rationalization behind some 40 features of the SA language , and shows how they enable rigorous communication which results frorn disciplined, recursive application of the SA maxim: "Every-thing worth saying about anything worth saying something about must be expressed in six or fewer pieces." Index Terms-Graphic language, hierarchic, requirements analysis, requirements definition, structured analysis (SA), structured programming , system analysis, system design, top-down. I. BLUEPRINT LANGUAGE N EITHER Watt's steam engine nor Whitney's standardized parts really started the Industrial Revolution, although each has been awarded that claim, in the past. The real start was the awakening of scientific and technological thoughts during the Renaissance, with the idea that the lawful behavior of nature can be understood, analyzed, and manipulated to accomplish useful ends. That idea itself, alone, was not enough, however, for not until the creation and evolution of blueprints was it possible to express exactly how power and parts were to be combined for each specific task at hand. Mechanical drawings and blueprints are not mere pictures, but a complete and rich language. In blueprint language, scientific, mathematical, and geometric formulations, notations , mensurations, and naming do not merely describe an object or process, they actually model it. Because of broad differences in subject, purpose, roles, and the needs of the people who use them, many forms of blueprint have evolved, but all rigorously present well structured information in understandable form. Failure to develop such a communication capability for data processing is due not merely to the diversity and complexity of the problems we tackle, but to the newness of our field. It has naturally taken time for us to escape from naive "program-ming by priesthood" to the more mature approaches, such as structured programming, language and database design, and software production methods. Still missing from this expanding repertoire of evidence of maturity, however, is the common thread that will allow all of the pieces to be tied together into a predictable and dependable approach. The author is with SofTech, Inc., Waltham, MA 02154. II. STRUCTURED ANALYSIS (SA) LANGUAGE It is the thesis of this paper that the language of structured analysis (SA), a new disciplined way of putting together old ideas, provides the evolutionary natural language appropriate to the needs of the computer field. SA is deceptively simple in its mechanics, which are few in number and have high mnemonic value, making the language easy and natural to use. Anybody can leam to read SA language with very little practice and will be able to understand the actual information content being conveyed by the graphical notation and the words of the language with ease and precision. But being a language with rigorously defined semantics, SA is a tough taskmaster. Not only do well conceived and well phrased thoughts come across concisely and with precision, but poorly conceived and poorly expressed thoughts also are recognized as such. This simply has to be a fact for any language whose primary accomplishment is valid communication of understanding. If both the bad and the good were not equally recognizable, the understanding itself would be incomplete. SA does the same for any problem chosen for analysis, for every natural language and every formal language are, by definition , included in SA. The only function ofSA is to bind up, structure, and communicate units of thought expressed in any, other chosen language. Synthesis is composition, analysis is decomposition. SA is structured decomposition, to enable structured synthesis to achieve a given end. The actual building-block elements of analysis or synthesis may be of any sort whatsoever. Pictures, words, expressions of any sort may be incorporated into and made a part of the structure. The facts about Structured Analysis are as follows. 1) It incorporates any other language; its scope is universal and unrestricted. 2) It is concerned only with the orderly and well-structured decomposition of the subject matter. 3) The decomposed units are sized to suit the modes of thinking and understanding of the intended audience. 4) Those units of understanding are expressed in a way that rigorously and precisely represents their interrelation. 5) This structured decomposition may be carried out to any required degree of depth, breadth, and scope while still maintaining all of the above properties. 6) Therefore, SA greatly increases both the quantity and quality of understanding that can be effectively and precisely communicated well beyond the limitations inherently imposed by the imbedded natural or formal language used to address the chosen subject matter.