Re: OO Design induces an existential crisis

Addendum to my reply:

> If I compare your principles to those described by Robert Martin,
> I'll go with the latter. They, at least, make sense.

Unfortunately, Robert Martin's principles are too vague. The duplicate
case/switch list "mental coupling" discussion and the
report-versus-calc-summing-loop discussion made that pretty clear I
beleive. I hold those two examples up as rough spots to be worked on in
his "coupling theory". (If you can clean them up, that would be great.)
His view of "coupling" seems clearly tied to estimated change path
probabilities, and I disagree with his frequency estimates in many
cases, based on my actual experience.

In short, Robert's principles are not the "magic clear-cut design math"
that some hope for or suggest exists in them. It is all right back to
human comfort/psychology and change direction probability estimates.
This is where most OO/p-r battles end up in the end.

Robert M. has some good ideas, but they need a lot of refining,
clearification, and explicit assumption statements.

-T-

.

Relevant Pages