That only means we can't get an absolutely optimal solution in polynomial time. Fortunately, we are able to get a solution arbitrarily close to optimal in polynomial time. Find the correct balance of time vs. optimality and BAM that NP-completeness isn't really a huge concern.

Now you're just putting words into his mouth. Nobody's expecting linear scaling, here! That is an entirely different question.

But its not just one codec that's involved, it's multiple. A typical video file will have at least two, one for video, one for audio. If you have alternate audio streams or subtitles, you need a codec for each of those as well. The purpose of a container is to let you take all these streams in their individual codecs and put them together in one file for easy playback. I'll grant you that it's not optimal to have some files with a given extension playable and some not, but with such a multitude of codecs in a single file it's just not possible to condense all that information down to a single file extension.

Absolute craziness is that I've heard the term used to refer to people in, say, Britain. I think Americans might be a bit overzealous in their attempts not to offend.

And has been replaced by a less precise word? I'm still curious what word I'm supposed to use to refer to the group containing my Siberian, Chinese, Pakistani, and Indian friends, since Asian seems to have been hijacked for no good raisin.

"We reserve the right to refuse service to anyone" and "if you return here you will be considered trespassing on our property" are all it really takes. And, for that matter, only one or the other is really necessary.

Wow I think I need sleep. I read that first line as "I haven't finished baking Apple pie." Can I have some? I'm kind of hungry.

I used to run Super Windows 64... I mean 64-bit XP, right up until the day I installed the 7 RC. Never had any issues with drivers, had very few software compatibility issues (all entirely with older games, not that 7 has been doing any better with them), and it generally ran like a dream. That said, 7 gave me a nice little performance boost (thank you hybrid SLI), so I'm happy with getting better performance on a shinier system.

But that would certainly not be the way that best suits you in this case.

Most of what you say is true. Fortunately, we are able to harness entropy sources to achieve truly random results. It's quite useful.

My brother prefers the order of start car, unlock doors, open door, get in, drive, so even your clever car analogy is bad. Then again, given that this is /., that's probably a given.

This was one of the very unintuitive things I fortunately learned early in my academic career, and it served me very well. Unless you have a ridiculously stringent time limit (as in you are expected to finish the problems no more than, say, a minute before time is up, if ever) you should always look at everything before actually doing any. Even when a few questions seem trivial. It seems silly and unintuitive, but in times when things are not as simple as they seem it will reward you immensely and in the times when things are legitimately simple you don't waste more than a couple of minutes (potentially less depending on test length) and make sure you don't run into any surprises.

Your interpretation really only makes sense if you assume that the steps must be done sequentially, which is usually never true for paper tests. Usually, questions are intended to be done in whatever order best suits you which, in this case, is the order that tells you to not do the rest.

I'm sure you've heard plenty about the relationship between correlation and causation, so I won't even bother repeating that unless you really want.

Though, I am interested in the chart comparing people who have used illegal drugs in the past year to people who have been drunk at least 51 days in a year, and noting that the only statistics that are higher for the illegal drug users are in drug possession or sale, which is rather obvious why, and arson, which is only by a tiny margin (read: statistically insignificant margin), yet I don't see anyone suggesting that every president we've ever had may be unfit because they've used alcohol.

Other than that, the only section that seems to be of any causal interest is about crimes to fuel a drug habit, an issue which would largely disappear with legalization, so it only counts for half credit at best.

Which completely and totally occurred in the cases of Obama and Bush? Or for that matter, most users? Hold them responsible for such actions. Even hold them responsible for doing things to increase the likelihood of such actions while under the influence (hence our drunk driving and public drunkenness laws), but until you show me that coke is going to result in such things, your argument doesn't hold.

Okay, let me rephrase: abusing power in order to violate personal privacy, which, while still ambiguous, is definitely a Bad Thing, perpetrated against another, unwilling person, whereas cocaine use is a potentially Bad Thing perpetrated against oneself.

Moreover, specificity is more or less irrelevant. Causing the death of a living creature is really vague, and covers everything from squishing an ant to harvesting a plant to murder, whereas engaging in respiratory functions is rather specific, but the first category is clearly worse than the second and it would be silly to suggest the reverse.