I read a book once that claimed that it was illogical to call oneself an atheist. The book claimed that to be a believer made sense: one could have had, say a personal revelation which convinced them beyond any doubt that God did exist. On the other hand, to believe with absolute certainty that something does not exist requires personal knowledge of everything there is, to be sure that whatever you're looking for isn't among them. The best you can say is that you doubt the existence of God, not that you know he doesn't exist. And therefore, according to this book, people who call themselves "atheists" are more properly agnostics (and the author says that atheists he's spoken to and explained this reasoning to are all quite happy to adjust their designations accordingly...)
I don't really buy it. It's an interesting point, that proving something's existence really is fundamentally different from proving its non-existence, but there's a lot of word-games going on here, and even beyond that it doesn't really hold water. After all, the author is willing to accept as a believer someone who is convinced to his own satisfaction that there is a God, even if said God hasn't paid him a personal visit. What's wrong with extending the same courtesy to atheists, and accepting that someone who is convinced to his own satisfaction that there is no God (not that he is not convinced that there is, that he is convinced there isn't) as an atheist?
Still, it's food for thought.
Added October 2, 2000: Upon reflection, I think the book's reasoning behind not calling someone an atheist who is convinced of atheism to his own satisfaction is that a believer may not have experienced a revelation himself, but could reasonably decide to trust someone else's word that he (the other person) experienced revelation, but an atheist would either have to know everything (and thus know everything isn't God), or be asked to believe that someone else knows everything (and that everything isn't God). It doesn't make it any more valid a point, to me.