Pseudonym: I don't know the answer to the NP-optimization question. Could you argue that in some cases it is not certifiable in P, since the only way to check would be to compute all of the solutions (NP task) and verify that the particular solution is optimal? Of course, I don't know if you can declare this for all NP-optimization problems, and it's making my head spin right now (I can only do this sort of math at certain times).

If anybody is curious about optimization problems, here is a Compendium of NP-Optimization Problems I found on the web.

Hacking

ghod: Hacking actually dates back to the early days of time-share systems at MIT. It was often used to denote crawling around in the innards of a program (eg, debugging code) or even a building (eg, exploring forgotten steam tunnels). It's still used in that way today at MIT (cf, The New Hacker's Dictionary). This is also where the phrase "a quick hack" comes from to indicate code thrown together to solve a problem quickly, but often in a confusing or inelegant way. No connotation of intent was implied, but it was generally done for good or at least neutral purposes. About 10-15 years ago, the meaning of the term began to shift to what it is today. Unfortunately, hacker has taken on a negative implication (because of a lot of bad behavior), causing people who don't hack maliciously to want a different term to describe themselves. Hence, "cracker".

Let's not get into the whole debate over "black hat" vs. "white hat" hackers, or even how much good a "white hat" hacking group like l0pht can do.