# Arlen Cox

Ramblings on my hobbies

## Unofficial Laws

As an undergraduate I had a friend named Chad who invented his own named “constant” that he called Chad’s “Constant”.  This particular constant is of use to any engineering or physics student where a numerical answer to a problem is desired.  Chad’s “Constant” is defined as the correct answer to a problem minus my answer to the problem.  Consider this example:

$P=A$

Problem $P$ is equal to some correct answer $A$.  However, I compute (possibly incorrectly) that:

$P=B$

To ensure that I get full credit for my assignment, I can always add Chad’s “Constant”, $K$ in order to get the right answer:

$B + K = A$

## Arlen’s Law of Performance

If I add this as the last step to any problem, I can never fail.

In a similar vein, I will define Arlen’s Law of Performance, which applies to software development.  It is this:

If there is any doubt that performance will suck, performance will suck.

It’s not quite Amdahl’s Law as there is no mathematics backing this up.  This is purely the observation that probabilistically speaking you are more likely to have bad performance than good.  If heuristics are used to accelerate a process there is always a care where they fail and that case will likely occur often.  SAT is an exception to this rule, but it has been carefully studied and a tremendous amount of effort has been put into making performance not suck most of the time.