A response to Systems Analysis Slipping Backwards . . .

Another view of Structured Analysis

Richard Kuper, President
R.L. KUPER, Inc. Management Consulting
July 18, 2006

In Systems Analysis Methodology Sliding Backwards you have defined "Structured Analysis" as "not a way of doing systems analysis, but just a set of conventions for documenting and presenting the results (or deliverables) of systems analysis." You have also implied that this method first came into being in 1978. Afraid I need to disagree with that, as I was doing precisely that professionally from 1973, although the term "structured" was not applied until around 1978, as you indicate. In fact, at a financial company where I was the manager of the programming staff in 1978 or thereabouts, the "new" concept of "structured" analysis and design, at least according to the manager of the group of "systems analysts" I was dealing with, was that the system specifications were to be written to such a level of detail that the programmers could practically code the programs from the system specifications. (As an aside, the concept of "structured programming" there was: no GoTo statements allowed!)

The concept that the system specifications were to be written to such a level of detail that the programmers were, in essence, nothing more than "keypunch operators"* created a major conflict at the time, as I started rejecting such program-level specifications. (A bit of background: Prior to my becoming the manager of the programming staff there had been a high level of turnover, in large part because the programmers were not given any opportunity to be part of the discussion or solution.) I am happy to say that I convinced my manager to back me up. System specifications went back to being system specifications. The programmers got the opportunity to create programming specifications and to have real conversation with the systems analysts. Not only did programmers stop leaving the company, the quality of the delivered product went up dramatically.

So, in the end, as I always try to ensure, common sense should always win out over whatever the latest buzz word or "methodology of the day" or "process of the day" might be.


* For readers not familiar with the term "keypunch operators": Before there was the Internet and before there was online programming, programs were created on machines that punched special holes (Hollerith codes) in paper cards based on what the keypunch operator typed, which, in turn, were read by card readers connected to mainframe computers.