Dear DS,
It greatly concerns me how the SQL language is inconsistent with its mathematical foundations (i.e. tuple calculus, relational algebra, set theory, so on and so forth); as a matter of fact, when doing complex queries and using operations like COALESCE, the basic assumption is that the answer being returned to you is wrong - and it is implicit that you have to "fix" the answer to be right. Why is this so?
Also, I've found a buffer overflow vulnerability and have managed to smash the stack in a popular Linux program; should I tell someone or try to hack a bank?
Sincerely,
Bang Splat