Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!wuarchive!psuvax1!psuvm!uh2
From: UH2@PSUVM.BITNET (Lee Sailer)
Newsgroups: comp.software-eng
Subject: Re: Programmer Licensing?
Message-ID: <89293.124011UH2@PSUVM.BITNET>
Date: 20 Oct 89 16:40:11 GMT
References: <39400056@m.cs.uiuc.edu> <242@cherry5.UUCP>
 <1989Oct19.210055.2770@mentor.com>
Organization: Penn State University
Lines: 41

The CACM occasionally publishes items they call "self tests".  Some guy
develops a sort of outline of some topic in the form of questions.  If you
can answer all the questions, then you are good at that topic.  Here might be
a mini-example:

1.  What is the primary output of the Analysis process?

2.  What is the difference between Design and Programming?

3.  Describe each of the following, and give conditions where each would
be particularly useful:

     Data Flow Diagrams
     Warnier-Orr Diagrams
     Entity-Relationship Modeling
     Structured Flow Charts
     Structure Charts
     Finite State Machines
     Structured English
     Tight English
     Literate Programming

4.  Give an outline of how computer programs might be mathematically proven
to be correct.

5.  How would you estimate how long it might take to create a system that
will have about 1,000,000 lines of 3GL code in the delivered product?

etc etc etc

Now, I don't claim this is a good test, just that it is an example of the
kinds of questions that might appear on one.  The curious thing about these
tests, is that there doesn't really have to be a rght or wrong answer to each
question.  One measure of the test taker's expertise is merely the proportion
of the questions that they feel they can answer.  These all happen to be
questions that I feel that I can answer, but you probably have questions that
you think *ANY* software designer ought to be able to answer, which I
couldn't.

In short, there doesn't have to be an agreed method for software engineering
before a reasonable software engineering test could be  constructeded.