[Haskell-cafe] Implementing PacMan

[Andrew Coppin]

What do we think of this, folks?

http://prog21.dadgum.com/23.html

(And the rest in the series, obviously.)

To me, it seems that this plan would *work*... but it wouldn't be very 
eligant. You'd have the code to respond to user input and move PacMan in 
one place, the code for collision detection in other place, the code 
that decides what to *do* about a collision somewhere else, the code for 
drawing the animations in yet another place... If you wanted to change 
one game element, you'd have to make changes scattered all over the 
place. The whole thing seems messy, non-modular and non-composible to my 
ears.

Thoughts, people?