mental models

[Vadim Nasardinov <address@hidden>]

 > Hmmm...I don't think I have a strategy, per se

I would contend that it is impossible to not have a strategy, although
it is very common to not be consciously aware of having one.  This is
especially true if we redefine the term "strategy" to mean a mental
"subroutine" that you use when approaching a particular task.

Although I don't have my copy of The Mythical Man-Month handy to back
up my recollection, I seem to remember Brooks sharing the following
interviewing technique.  When he interviewed programmers applying for
a job, he would often ask the interviwee something like, What is your
model of a calendar?  Most people would not even understand the
question. Others knew exactly what he was talking about and could
describe their mental model of the calendar in great detail.

Assuming I haven't made up this quote, I think Brooks' point was that
programming is a highly visual activity and those who are good at
representing things visually in their mind are likely to be good at
programming. However, different people have different ways of
visualizing stuff. For example, you could represent the 12-month
calendar as a stack of 12 sheets, each having the familiar breakdown
of one week per row, for a total of four to five rows. Or you could
represent the calendar as a ring with black and white segments where
black segments represent 31-day months, and white segments represent
shorter months.  Or you could represent it as two concentric rings: the
inner one with 12 segments and the outer one with 52.

Not all people operate primarily in the visual domain. If you give
your friend directions to the Microcenter store, she may remember your
directions auditorially: ".. and then he said, take a left onto
Prospect and go towards the river, and then he said, take a left onto
Putnam at the light...'  Or she may draw a map in her head and forget
the sound of your exact words, once the image is committed to her
memory.

The professor who taught the group theory class back at college once
remarked that his way of performing (symbolic) computations was to
mentally paint the terms of an expression black and white. As he
performed calculations in his mind, the respective number of black and
white terms would change. If he was getting more white than black, he
knew he was on the right track. (Those math people can be pretty
wacky.) The remarkable thing here is not the particular strategy that
he used, but rather the fact that he was conscious of it at all.

These processes are largely unconscious and we aren't aware of them
most of the time. You would have to literally slow down your thinking
to start noticing your thought patterns.  And it's really hard to
think slowly.

Back to my point. You cannot not have a strategy in the sense that I
just described. Granted, that's not what my original post was about,
but I hope the connection between these two definitions of the notion
of "strategy" is obvious.  Ultimately, it would be extremely
interesting to figure out how, say, Paul Graham visualizes programs in
his mind's eye.  Unfortunately, this kind of thing is next to
impossible to figure out over email.