Heuristic is....

"A fallible method of solving a problem or making a decision."

There are only two key ideas that make a heuristic: 1. It may help solve problemor make a decision. 2. It may not.

I consider a test idea to be any idea related to a test you might want to perform.

A heuristic (as a noun) is "a(fallible) method for solving a problem or making a decision". Some peoplealso say "guideline" or "rule of thumb". When used as an adjective (e.g. "aheuristic approach"), "heuristic" means "conducive to learning (albeitfallible)". Contrast "heuristics" with "algorithms", which I'll define hereas "a step-by-step procedures or recipes for solving one particular class ofproblem".

Heuristics are
- not to be confused with rules or edicts
- context
-dependent, situational
- potentially contradicted by other heuristics
- to be used well, need to be used by someone with the appropriate level ofjudgement and skill to choose and use them appropriately- not to be considered perfect, but good enough.
- typically "fast and frugal"-
-much faster than more rigourous and complexforms of analysis, plus
- according to Gigerenzer, often more accurate than other rigourous andcomplex forms of analysis

The hidden secret is that /all/ of our methods of solving problems or making decisions are heuristic--good-enough approximations. Even mathematics orphysics are fallible outside of a context in which (There's a wonderful linefrom "An Introduction To General Systems Thinking", by Weinberg, that goessomething like: "Mechanics, then, is the study of things for which theapproximations of mechanics work sufficiently well.")

If you want to understand heuristics, I STRONGLY recommend GerdGigernenzer's recent book "Gut Feelings", or George Polya'snot-at-all-recent "How to Solve It", or Billy Vaughan Koen's recent butnow-hard-to-find book "Discussions of the Method." They're all verydifferent books, in fact, so maybe that "or" should be an "and".

(ideas stolen from James Bach, Cem Kaner and Michael bolton)