4 ms·
This seems really stupid to me. An action has defined effects, there is no reason to think of all possible nonexistent side effects. Then they allow for that as
by phkahler 9y ago
This seems really stupid to me. An action has defined effects, there is no reason to think of all possible nonexistent side effects. Then they allow for that as a default and point out an apparent exception: what if an object is moved into a can of paint... This is so incredibly stupid I can't imagine people take this seriously.
- danharaj 9y agoThe first interpretation of the frame problem is about mathematical entailment. If it isn't stated that a particular predicate on situations is preserved by taking actions, then it does not follow that the state is preserved. If you want to reason about situations in first order logic you need to pin the given theory down so that non-effects become deducible from the axioms. Either you can introduce frame axioms, which are tedious and slow, or you can try to solve the problem in a more sophisticated way. It's definitely not a trivial problem. A problem about logic in the narrow technical sense often has analogies in less technical philosophical considerations. That's what the rest of the article is about. You're taking too much for granted about how to formalize reasoning to much how humans intuitively go about it.
- lisper 9y ago> there is no reason to think of all possible nonexistent side effects But there is. Actually there are two reasons. The first is that the "non-existent side-effects" are only non-existent in certain circumstances, i.e. an action's effects depends heavily on the context in which it is executed. Moving X from A to B can have lots of other effects beyond changing the location of X depending on the circumstances. These effects can be arbitrary -- consider for example the case where X is a switch and A is the OFF position and B is the ON position. The second is that there really is no such thing as a truly non-existent side-effect because even a null action -- waiting, doing nothing -- results in change, and sometimes that change turns out to be the best way to achieve your goal. For example, if you're a farmer whose goal is to water their crops, one way to do it is to do nothing and wait for rain. But to figure out whether or not that's what you should do you have to know an awful lot about how the world works. In fact, waiting and doing nothing is often an integral part of many goal-achieving procedures. Think about, for example, traveling as a passenger on an airplane. The usual procedure goes something like: Buy a ticket Wait Go to the airport Wait Board the plane Wait Disembark the plane If you leave out the null actions, that procedure fails badly.