Having a boy and a girl is twice as likely as having two boys. You are right when you say that the son being talked about could be the older or the younger: this is taken into account in the calculation P(X|boyboy) = 1/7 + (6/7*1/7) = 13/49, which says that either the first boy was born on a tuesday (1/7), in which case we need not look at the second, or he wasn't (6/7), in which case we check whether the second was (1/7). This covers all cases where "I have a son who was born on a Tuesday", without counting twice the case where I have two sons who were both born on Tuesday.

There are no games with words: the problem statement is perfectly clear if you read it carefully.
We are only separating the cases by order of birth to help you understand how the probabilities work out. We could just as well say that there are three cases:

1. "two boys", with probability 1/4
2. "two girls", with probability 1/4
3. "a boy and a girl", with probability 1/2

The last case is twice as likely as either of the others because the two births are independent events (by assumption; in real life there is probably a slight correlation, girl births are more likely than boy births, etc., but it doesn't deviate very far from that).

If you're still unconvinced, look at this: http://codepad.org/kMVsqzyT

If you write it that way, then the probabilities are 1/6, 1/6, 1/3, 1/3 respectively. It is, in fact, twice as likely to have a boy and a girl (in whatever order) than it is to have two boys, no matter what you call them.

To clarify: I mean, of course, that the fact that the weekday of birth is specified is significant. It doesn't matter whether that day is Tuesday, Friday, or even Monday. (Assuming uniform probability of childbirth across weekdays, etc.)

Yes, there is an assumption that the weekday of birth of a child is independent from that of its siblings, and uniformly distributed. The twin thing is only one of many factors that can violate one or both of these assumptions.

Let's assume that, if I have two children, it is equally probable that they are born as boy+boy, boy+girl, girl+boy, girl+girl. If I have one boy, girl+girl has probability 0, and the other options are equally likely, so they have probability 1/3. If it is known that I have a boy, there is 1/3 probability that the other is also a boy.

X=one boy is born on a tuesday
P(X|boyboy) = 1/7 + (6/7*1/7) = 13/49
P(X|boygirl) = 1/7
P(X|girlboy) = 1/7
P(boyboy) = P(boygirl) = P(girlboy) = 1/3
P(X) = (1/7 + 1/7 + 13/49)/3 = 9/49
Using Bayes's theorem:
P(boyboy|X) = P(X|boyboy)*P(boyboy)/P(X) = 13/49 * 1/3 * 49/9 = 13/27

Which is different from 1/3. So yes, the weekday of birth is significant.

"Track down political adversaries"?? What the hell is that supposed to mean? There are parties, and newspapers, and trade unions, and countless public personalities that daily denounce Berlusconi as the devil's incarnation, and you seriously think he's drafting elaborate plans to find out the real identity and whereabouts of the freedom fighter who wrote "belrusconi scemo!!!11!" under the pen name of goku92 on a videogames forum?

Yes, it must have been a fluke. I tried it out with an article on Ars Technica, and indeed it fetched all pages immediately.

You know, I used the reader feature for the first time while reading TFA, just to piss him off. But it doesn't look that useful to me. It doesn't start loading the next page until you scroll down to it, so you still have to stop and wait in the middle of your reading (unless you get in the habit of doing a quick scroll to the bottom in advance).
Also, there is no way of knowing what is being left out of the display, either by design or due to a parsing bug. How do I know that I'm not missing a paragraph or a sidenote? I think I would only use this feature on sites with extremely annoying designs, where the usability gain overrides those concerns. I think the best countermeasure for concerned webmasters is simply making sure their websites don't suck.

That was Clinton's big idea back when he promoted China's entry into WTO, wasn't it? But what actually happened is that they just got rich, yet they are not any more democratic than before. I think they did get more nationalistic, though, so that's something.

I'm not surprised. If I had a million kabillion dollars, I'd hire my own scribe too.

I think you can use Jack OS X.

There is.

That's a good point. Someone would have to look through the various features to see what their status is (draft, approved...). At any rate, it's true that this is a showcase of "what we would like HTML5 to be", rather than "what it is".
On the other hand, it's quite unlikely that people are going to start building websites that rely on those features, given Safari's small market share. What might happen is that web designers get interested and ask other browser makers to hurry up and add support for Apple's flashy stuff. You might say that's a bit underhanded, but it probably won't be a bad thing in the end.

HTML5 is still a work in progress. They could have made a demo that only uses those features which are already widely supported, but it wouldn't have been as impressive. Or they could have made a demo that uses the latest bleeding-edge proposals for HTML5, and let it fail on most people's browsers - perhaps even worse.
Given that it's meant to be a showcase of things to come, it makes sense to require you to use the one browser that currently works with it. Even Mozilla sometimes releases demos that require the latest Firefox beta to test. Using browser sniffing to enforce it is certainly bad form, but they probably thought that otherwise people would just click through, see a broken demo, and not even realize they aren't seeing what they're meant to see. Hopefully they'll relax the restriction once (if) more browsers implement support for these proposed new features.