Re:Automating Formal Proofs for Reactive Systems
by taeric on Sunday June 22, 2014 @01:43PM
(#7928569)
First, "no bad things can happen" is a very nebulous requirement. And you eventually just get into a "who watches the watchmen" kind of game.
I think the problem with a lot of ideas around making a "safe" algorithm for money matters, is that you basically just wind up defining the game that your "opponents" will now be playing against. That is, if you have your rules in the open, but they don't, then they are the ones controlling the game.
Consider, it is relatively easy to make money off of a casino. They just don't let you use the strategies that work. Which is fine if you can just boot off actors that you think aren't "playing well." Not so easy otherwise.
I think the problem with a lot of ideas around making a "safe" algorithm for money matters, is that you basically just wind up defining the game that your "opponents" will now be playing against. That is, if you have your rules in the open, but they don't, then they are the ones controlling the game.
Consider, it is relatively easy to make money off of a casino. They just don't let you use the strategies that work. Which is fine if you can just boot off actors that you think aren't "playing well." Not so easy otherwise.