Dan Shue (dshue)
CS 190 - Spring 2003
Mythical Man-Month Essay


In "The Mythical Man-Month", Frederick P. Brooks writes about some fundamental
software engineering practices that are clearly applicable to CS190.   I wish
to address 2 of them: conceptual integrity and extensive testing.

Brooks starts off Chapter 4 by saying that "conceptual integrity is the most
important consideration in system design" (p. 42).  He argues for a system that
is unified under one set of ideas rather than a system that has several sets of
independent, uncoordinated ideas.  Therefore, the design process must come from
"one mind, or from a very small number of agreeing resonant minds" (p. 44).

At first, I thought that this wasn't the best way to go about it, because in my
mind, having more people brainstorming for ideas yields more productive
results.  In regards to CS 190, I thought that the requirements and
specifications stages could have been done in groups, but now I see why we're
doing them individually.  If multiple people worked on the requirements and
specifications, our current projects would have been a lot less clear and
defined, making it harder for us to implement later on in the semester.

Another important idea that Brooks mentions throughout his book is the concept
of extensive testing, which is undoubtedly the most underestimated stage in any
given project; in fact, Brooks argues that 50% of the time spent on a project
ought to be devoted to testing and debugging.  I read this before I started my
CS 190 project, but because of my laziness and optimism, I didn't pay too much
attention to testing while I was writing up my requirements and specifications.
Hopefully, I'll be able to write up more specific testing procedures by the
time we have to test our projects; otherwise, I'll have to suffer the
consequences of not thinking ahead.