When should you resign?

The post 2 slots below (Some Brotherly Love) provoked some correspondence. The question was "Why did black resign?" and ignoring issues about losing to your brother etc it is quite a good question.
However I'll answer it as a slightly different question, "When should you resign?"
There is a quote attributed to the Danish GM Bent Larsen which is "You should resign when the spectators understand why you are resigning". Of course this is then dependent upon the chess knowledge of the spectators, and is therefore hard to judge.
For most chess players I think the unwritten rule is "You resign when you believe your opponent is going to win". Of course taken to an extreme that may lead to more games like Fischer v Panno 1970 (1.c4 Black resigns), but usually it is the position on the board, rather than the opponent you face that counts.
Some players use a material rule of thumb to decide when to resign. One member of the PNG Olympiad team says he won't resign when down a piece as his opponent may blunder it back. The obvious rejoinder to this is "But which player has done the blundering so far?"
Interestingly enough these unwritten rules only seem to apply in "normal" situations. I've often seen (and heard of) players playing on after losing a piece before move 10, when if they had lost a piece after say move 15, they would have resigned instantly. It is as though the course of the game isn't an "usual" one, (ie it isn't usual to blunder so early), and therefore "usual" guidelines do not apply.
And one final thought. I once heard a quote concerning chess as a spectator sport for non-serious chess players. "They play for 4 hours and just when it is getting interesting, one of them gives up". Clearly for serious chess players the interest is what happens before the resignation, for casual players it may be in how the game is finished off.
Maybe it is time to remove the option to resign from the laws of chess.