SQL Server Query Performance Tuning Distilled, by Sajal Dam

This book doesn't teach you what to put in each table and how to put them together. But, when you have large tables it's very important to have good indexes, and to know how to evaluate that your queries are using them correctly. There is a good explanation of using the Profiler here, as well as a nice tutorial-style approach which explains how to improve your indexes based on the execution plans that the Query Compiler spits out.

When you divide the problem that way, you get the two cases which equal likelyhood:

in case mentioned boy is older:
B, B p = 1/3
B, G p = 2/3

in case mentioned boy is younger:
B, B p = 1/3
G, B p = 2/3

Why is it in both cases 1/3 vs. 2/3, and not, as you claim, 1/2. You have to look at the probability distribution of the underlying set. It is:
B, B = 1/4
B, G = 1/4
G, B = 1/4
G, G = 1/4

So all have the same likelyhood. In our breakdown into the two paths "younger and older" above notice how "B, B" is mentioned in both of them. Since "B, B" will occur on average just as often as "B, G" or "G, B", all the "B, B" cases will be split between the "boy younger" and "boy older" path, and it's probability of occurring in either path is thus halved when compared to it's alternative

When all Probabilitys are added up, the end result for the probability that you get B, B is the same.
p = 1/2 * p[path 1] + 1/2 * p[path 2]
p = 1/2 * 1/3 + 1/2 * 1/3 = 1/3

No, he didn't "see" a random memory error. Maybe you can see them, i sure as hell can not.

He noticed a program had stopped working for unexplained reasons, and did some pretty nifty debugging to get to the bottom of it. Yes, i'd have rebooted too. But not because i think that what he did was a waste of time, just wouldn't have known how to do it.

And as for memtest, yes you can use it to find out if you have really shitty ram. But it's no magic wand, what if the ram is working mostly fine, and an error appears on average only once a year? You'd have to run memtest for years to test for that with certainity, and what's the point of having a computer then if you can't use it?

What i found funny btw is the hostname "psychotique", no wonder it is giving problems.

I was actually kind of puzzled why google took out that feature. Was allways so funny to read "swim across the atlantic ocean..". Maybe their lawyers thought about it and decided to not take any chances :(

Replace "copyright protection" by "to use their exclusive rights granted by a copyright to prevent others from selling copies of their product". I don't know if this definition is technically correct, but from the context of her speech i think it was what she actually meant. Maybe it was prone for misunderstanding, that some people might think that OSS is PD, but i think your categorical "False" is too harsh a judgement.

Medicine is heavily regulated, no free market there

Of course, if you remove those regulations you get tons of quacks ripping people of. So they are propably there for a good reason. But when it takes millions and millions just to get the paperwork done to start some trials, it seems the bureaucracy is overblown currently

If you ever look at :version, it spits out this huge list of features, each with a "+/-" in front.

I suppose when you build your own custom verion, there is a way that you can configure your version to leave most of those features out.

In fact debian seems to have done just that, they ship "vim.tiny" in the base install with just 640k, which should be enough for everyone.

It's not exactly "real time" if you can build an oil refinery in less than the time it takes for your units to cross the map.

Perhaps you have underestimated how awesomely large my map is.

1) they still showed it - It was not censored, the fine was after the fact

2) you can see far more explicit content on US TV (Cable)

All the hand-wringing, apologies, and remedies to stop it happening again were done by the TV company before the FCC were involved, the fine was incidental

Do you really think the FCC would or could shut down a national network TV station?

What's a niched product? Seems I spend too much time mainstreaming...

SELinux / AppArmor, but more so, then ?

repeat 5 times

The previous gen. NV Cards don't do DX11, thus where running this bench in DX10 mode. Which is kind of a misleading thing for THG to do, while they explained it in the text they should have made seperate chart for those two modes to make it clear on thirst glance. There is a comparison of image quality in Unigine Benchark (DX10 vs. DX11) out there somewhere, and the difference is night and day.

In civ4, catapults do only damage part of the units in a stack.
So if i send a stack of of 30 Units on attack, and AI smashes a Catapult against that, only a few units will get damaged. A sizeable amount of catapults you have to worry, put when there just a few you can in practise just ignore them mostly.
The only thing cats are really strong at is in city-attack, because they can bombard defences and you can promote them to get a bonus at it.

They could of course change the game-balance so that catapults would damage a stack more, but then there would be crys that catapults are overpowered and need to be nerfed. Better to do away with the "50 units on same square" silliness.