i.e. a 6 dB reduction in overall Sound Pressure Level. Just enough to be noticeable, but not really a huge change.

The real issue is that no one has reliably determined what noise levels are "acceptable" for sonic booms. A lot of recent work by NASA and JAXA has focused on which types of noise metrics correlate best with annoyance (in short, Perceived Noise Level seems to do about as well as anything), but it's the politicians who have to decide what levels are allowable for overland flight. For that, you have to identify their constituents that care. Quite honestly, the big boys in the aviation industry aren't that interested in supersonic flight--few will pay much more for their airplane ticket to save a few minutes of flight time, especially with so much extra time now spent in security. In fact, airliners have slowed down in recent years to save fuel and lower ticket prices. The guys pushing for a rules change are the business jet manufacturers, like Gulfstream, who do have customers who care about speed (and can skip the lines at the airports). How much noise are people in "flyover country" willing to accept so that rich CEOs can get from NY to LA an hour sooner?

It does! I have an IBM Model M attached to my Windows 8 PC, which gave me a brief bit of trouble until I learned that I could use Ctrl-Esc instead.

It's not the most ergonomic key chord, but as an Emacs user, I don't care about that anyway!

I'm also literally connected to Facebook (using jabber-mode in Emacs) at this very instant. If they've had XMPP shut down for over a week, they sure did a shitty job of it.

This isn't really true; the Constitution was originally silent on the matter of who could or could not vote, leaving it up to the legislatures of each state to determine the specifics of elections. Moreover, only the House of Representatives require direct election by "the People"--State Legislatures could choose Senators and even Presidential Electors as they saw fit. Voting rights in the US were very much non-uniform; while most states initially limited the franchise to white property owning males, New Jersey allowed property-holding women to vote from the start, some others jurisdictions allowed non-whites to vote, and of those a handful allowed freed slaves to vote. It wasn't until after the Civil War that uniform standards started to come into place. And we can all be thankful for that--how could one possibly believe the country was better governed when women, blacks, natives, and more were denied a voice? Your enlightened government of the past allowed one man to own another, said women were no better than slaves to their husbands and fathers, and coldly committed genocide against natives who had the temerity to live on land white property owners had their eyes on.

The US prior to the civil war had great ideals, but fell woefully short of them in practice as the white property-holding class ignored the rights of others whenever they stood in their way of their economic interests. Everyone wants something for nothing--I'd rather not go back to the time when only a few got it, at the expense of everyone else.

Exactly right. Sony's gaming division is about ~10-15% of the company by revenue; how do you "spin off" 85-90% of the company? While I think the article has a point that Sony has too often crippled its products in order to help other business divisions succeed, I think the solution is more of a refocusing of corporate priorities than dissassembling the company. For instance, there are a lot of shared technologies that go into the range of consumer electronics products Sony makes, and fairly obvious advantages to sharing the IP and manufacturing capability between business units. Moreover, there's obvious places for cross-promotion of Sony Entertainment's products with Sony consumer electronics and games; but it's only going to be a good strategy in the long term if consumers feel they're getting something extra and not just having Sony's marketing pushed down their throats. Sure, Sony's reputation has, overall, been on the decline since the 80s, but with the right management priorities, I think they could easily be the Japanese equivalent to Apple. They just need to focus first and foremost on making products consumers will want to buy, and stop getting lost in protecting the existing business where it may not be so successful.

Aaaaaaah! How hard is it to include "stdio.h" from the get go? "Trivial" warnings are usually trivial to fix--and they should be fixed so that the serious warnings stand out. If you let your code develop with all warnings suppressed, than I can see how you can end up in the position where turning them on leads to hundreds of "insignificant" warnings appearing. Don't let that happen. The compiler writers and designers of the language thought long and hard about what should and should not be a warning--and most warnings are indicative of a problem that had ought to be fixed, even if it doesn't happen to cause a "significant" problem on this platform, on this data, and in this context. If you start out with warnings enabled, you learn pretty quickly how to write warning free code and it never becomes a problem. If you don't; well, then you end up with the completely avoidable and entirely self-made problem you describe!

Or you could actually look up the meanings of those words and discover that:

A democracy is simply a government where power is ultimately wielded by the people who are governed, whether directly or indirectly.
A republic is merely a government without a king or other hereditary ruler.
A dictatorship is a government ruled by a dictator, who singularly holds absolute power over the government.

These definitions are clearly not mutually exclusive--a democracy is often a republic (the US) but sometimes not (the UK). A dictatorship is often a republic (the Syrian Arab Republic), but is sometimes not (the Kingdom of Saudi Arabia). However, a dictatorship is never a democracy, except, I suppose, in the unlikely case of a country populated by a single person!

Most modern smartphones do contain GPS, as it is required for E-911 on some networks (namely AT&T) and is useful in many cases even on other networks. Most cell towers do contain actual GPS receivers, as well--not for their own positioning, since that obviously doesn't change much, but for accurate time and to bootstrap GPS equipped phones with the local almanac and ephemeris information so they don't have to take a few minutes to get a location fix. It's true that many phones will prefer radio-location, because it's faster and works better inside, but actual GPS, especially bootstrapped by nearby towers, is extremely common these days and virtually every smart phone has it. iPhones, for instance, have had actual GPS receivers since the 3G. The newer ones even support GLONASS. I was just in a really remote mountain with absolutely no cell reception and had no issue getting an accurate fix on my phone--it just took the phone a couple minutes to figure out where it was.

Conversely, US Government travel is always economy, and government contract fare seats are often the least desirable seats available (rear-rows, bulkheads, and middle). Although, rental cars are usually intermediate, not economy, and hotel per diem usually lets you stay somewhere pretty decent, though never luxurious. Altogether it's actually pretty reasonable for the tax payer and the traveler--however, the paperwork is still a nightmare!

I worked for a major automotive component supplier who designs and builds parts for most of the major automakers. I wanted to make a few of relevant observations from my experience.

First, all parts were extensively tested for function and safety. Designing a good test that is representative of years of field use is very difficult, but none of the automakers seemed lax in their testing requirements. Some were pretty quick to dump performance for a cost savings, but I don't know any who were, these days, willing to sacrifice on reliability. There weren't many arguments with customers about the cost of testing, and it was generally thought that some tests demanded by OEMs were needless, but we'd gladly take their money anyway.

Second, parts were regularly improved based on analysis of returned parts. The best source of these were fleet vehicles, which provided lots of high mileage parts back to the OEM--each and every one of these returns was examined, graded (often by some poor intern), then archived for future reference should a problem develop. I remember one incident where some tiny steel spring clip broke--this had never been seen before, so the entire engineering department was re-directed to determine the cause. Thousands of old parts were pulled out of storage and re-examined. I don't think we found another broken clip, but it was a big deal.

Lastly, parts were frequently revised for better performance, lower cost, or better reliability. Little bits and bobs, like switches, valves, fasteners, connectors, etc., were often used on numerous vehicles by a number of manufacturers. Each part had at least two sets of drawings and part numbers. One set was for our use, as the supplier, and had every detail labeled. Another drawing was prepared for the automaker, with only the details relevant to them called out explicitly. It was, in a sense, an engineering contract--we'd agreed to provide everything as described on that drawing as the same part number, but were free to change things not called out. Once I pulled up about thirty drawings produced for the same part, a tiny thing used in many of our products, to see whether we could change the part to an improved steel that was cheaper and tougher for this application. In all of the automaker drawings, the material spec was loose enough for us to change without asking for a change in the drawing. Our internal part number did change, but as far as they automakers were concerned, they were still using the same part.

Anyway, it's quite possible that someone might make a fix to the ignition switch without GM even knowing, and certainly without requiring a change in part number. In my experience, all of the majors are actually pretty good about testing everything and they all really do want to sell people reliable cars, as even the US big three have come to realize that each lemon they put out there can sour a family of customers on their cars for life. Management can be boneheaded about a lot of things, but I really don't think this is one of them. 100% safety isn't possible, no matter how much is spent--but they all get pretty close. Just look at how the fatality rate has plummeted over the last few decades, despite more traffic and more collisions.

That list omits a few more or less independent automarkers. For US market examples, where are Mazda, Subaru, Mitsubishi, Suzuki and Chrysler?

Except, obviously, not everyone survived. Car accidents are one of the leading causes of death and injury in the developed world. These features (e.g. ABS, ESC) have been studied many times, and almost universally demonstrated to significantly improve driver safety. As safety features like these have been adopted by automakers, by choice or by regulation, overall automotive fatalities rates have indeed decreased. There is virtually no actual evidence to the contrary--just anecdotes from drivers who believe, in the face of the evidence, that they are too good for driver aids.

I'm all for giving the driver the choice to turn these features off for some fun at the track, but I'm fine with them being universally installed in new roadgoing vehicles.

Many people make this mistake. The "morning after pill," e.g. Plan-B, is simply an extra-large dose of hormonal birth-control, and is indeed emergency contraception and not an abortifacient. RU-486 is an "abortion pill," but it's typically taken after several weeks to induce an abortion after the fertilized egg has been implanted and begins to develop, not within 24 hours (before implantation). According to wikipedia, RU-486 could be used in a much smaller dose as emergency contraception, but in practice is only used for such in China and Russia. Confusion between the two pills is often used in the US to rally opposition to OTC availability of Plan B, which is safer than OTC painkillers.

While bird strikes can happen at 10,000ft+, they occur with much much higher frequency near takeoff and landing where airspeeds are lower. Plus bird strikes can lead to disastrous consequences when the occur on other parts of the aircraft, such as engines or control surfaces, not just the canopy. Is there any evidence to suggest the track record of the T-38 is significantly worse than the rest of the Air Force fleet? There are finite resources to marshal and a great many things that could be improved in across the range of USAF aircraft--without any context, it's impossible to know whether fixing this one vulnerability is worthwhile.

RDRAND may or may not be compromised--without seeing the implementation, it's completely impossible to know. It's very much possible to create a stream that looks like perfect random "noise," but is actually deterministic; that's the whole idea behind encryption, after all. That said, Linus is also right that even if it is compromised, it doesn't hurt the entropy of the random number pool to include it. It would still strengthen the apparent entropy of the pool against anyone who did not possess the "secret" behind RDRAND, and would not weaken the entropy pool against those who did. People are right to be paranoid these days, but there's nothing to be gained by keeping the hardware random number generator out of the entropy pool.