There has been recent debate over whether a failing action should be illegal or not. Here's some arguments.
First, let's take a look at how performing actions by announcement works. You write a message stating that you perform an action, and somehow, when you send off the message, it happens. (Note that this is actually ISTID, except it's rule-sanctioned.) Now, the announcement that performs an action is obviously a statement. But it does not seem to imply that anything is true or false - at a stretch, we can say that it states that the action it purports to perform is successful. But I think this is too much of a stretch - it does not in fact say that and it's quite a leap to infer that from the action. Let's look at announcements another way. What if, instead of something that triggers an action, it's actually the action itself somehow? Personally I think that this perspective doesn't really make all that much sense (I sure as heck don't understand it) so I won't address it. Anyway, all this goes to show is that it's basically impossible to assign reasonable truth values to actions and even if you do it you need a large amount of inferrence (too large, in my opinion). Also note that if an action is not performed by announcement, it does not have a statement attached to it and therefore can fail without consequence by some recent (proto-)proposals. I'd also argue that failing actions/announcements causing failing announcements being illegal, even if the performer thinks they'll fail, is decidedly non-Agoran. The identity problem, however, is a large one. I support (I think it was) Goethe's (proto-?)proposal to make being misleading against your identity illegal. (I also second ais523's arguments relating to the PNP's voting.) tusho
