Path: utzoo!mnetor!uunet!lll-winken!lll-lcc!ames!sgi!wdl1!rhj
From: rhj@wdl1.UUCP (Bob Jones)
Newsgroups: comp.software-eng
Subject: Re: What is CASE (long)
Message-ID: <3850006@wdl1.UUCP>
Date: 5 Apr 88 17:38:55 GMT
References: <18625@think.UUCP>
Lines: 29

Some added comments on the general notion of CASE tools:

The various design methodologies are simply techniques for viewing a
problem.  To the extent that a methodology helps you understand the
problem, it is good.  If all you use it to do is document the fact
that you don't understand the problem, then it is a useless exercise.

The various toolkits are just automated reflections of the methodology.
I would agree with the analogy that they are just glorified word processors
except that the good ones do have the ability to verify that the model
is correct to the extent that it conforms to the rules of the methodology.
This certainly does not assure the best design or even a good design.

The real key to the proper use of these tools is the understanding of
the methodologies. Without proper training, the results will be sporadic
at best.  

The old phrase "there are many ways to skin a cat" applies here.  There 
are many ways that a software analyst/designer/developer can approach
the problem of understanding what the target system should do. Data flow
diagrams are one way (i.e. one mothodology), but there are others.  The
good news about Data Flow diagrams is that a lot of organizations have
used them and found that they do help, if used properly by trained people.
They are, therefore, a somewhat "proven commodity". The bad news is that
there are no really definative and objective studies that I am aware of
that clearly show just how much better Data Flow diagrams are over various
alternatives.  If anyone knows of any such studies, please let me know.

Bob Jones
Ford Aerospace