to code or not to code?

[David Sinck]

\_ It's best if you liberally intersperse code with comments, explaining not 
\_ only what is happening and why, but also why you *didn't* do it some other 
\_ way.

For a long time I've maintained that I want to find a book of what
*not* to do...the Tome of Anti Knowledge.

"Don't do a multi-dimensional-mip-map because the computers can't
support that much memory."

"Don't issue rm -rf /" 

and so on.  All the academic papers out there explain how they've got
cool results.  How many have you seen that explain, "well, this petri
dish is dead and empty when it should be at least growing."  

David