Obviously, depending what you do with the information and what the originating sites use policy is it may be legal (for instance google cache). You will be diving into shark infested waters when you start charging or building the scrapes into some sort of product. And always, even if you are doing something that may be perfectly legal, if someone sues you for it, do you have the resources (time and money) to fight the battle? Better to invest in a lawyer now and get reasonable advice than try to gather it from perlmonks. =)