I was reading "The mythical man-month" only this weekend, which starts with the observation that "everyone knows" that two kids in a garage can do more than a corporate development team, and then points out that, if this was actually true without caveats, corporations would hire two kids in a garage every time. There's a difference between producing a standalone program and developing/maintaining a product system.

If you want to organise your address book with a joystick or find the nearest restaurant to your television, not at all. But if we're talking about games, I don't think there are "millions" of great games for Android. ISPs who already resell TV channels may have one or two ideas about how to licence games. For example, you don't need to read French to spot the logos and brands on this ISP's website

In France, where almost all domestic broadband is "triple play" (phone, TV and Internet), at least two of the major ISPs offers gaming as part of the functionality of their latest glorified router package. You can't get much easier to install than "It's already there", and the ISPs already have a distribution model that they use to sell view-on-demand video.

The assumption in TFA seems to be that PHP does something that couldn't be done otherwise now, or does it more easily, or something. But I don't think that's true. There are alternative languages and alternative ecosystems now. IMO, most PHP coders don't use PHP because they have looked at the alternatives and decided PHP is the best choice, or the least bad choice, or any other choice. People use PHP because either

1: It was forced upon them for some reason or

2: It was the first thing they found and it was good enough

If that's the case, it's irrelevant how fantastic the alternatives are.

Also, while PHP code can be truly terrible, people who are determined to write terrible code will do so whatever the tool. You can use almost anything as a hammer if you try hard enough. The myth that The One Right Language somehow makes bad programmers good is still alive, especially within the Python community, but it's stll a myth.

When good programmers have no choice but to use PHP, they'll find a way to build something that is workmanlike even if it isn't beautiful. When bad programmers program, the result is going to be bad regardless of the language.

That is a truly terrible article.

To summarise the logic of TFA, America doesn't do well at standardised international tests, but the average level of scientific education is clearly dazzling because of Silicon Valley in general and the iPad in particular. Except that, last time I looked, the iPad was using a processor core designed in Cambridge, UK. I believe there are one or two other foreign contributions to Silicon Valley.

And, even if that were not the case, what percentage of people resident in the US work in Silicon Valley?!

My impression of US education is that, like many other aspects of America, the bell curve is very wide. The best of US education is possibly the best in the world. The worst is very bad. So, if "American education" means "the best of American education", there's nothing to worry about. If it means "what 90% of Americans have understood about science", or even "median American comprehension of science", the answer might be different. Or not. But the quality of science is not about whether or not a huge multinational with most of its labour outsourced can ship a commercial product.

(FWIW, I'm writing this in France, and I don't get the impression that science teaching here is great either.)

Both unsecured connections I can see here in France are from ISP-provided routers - one of them is mine. This is how French ISPs provide roaming wifi for their clients - leach a bit of bandwidth of domestic connections and make it available via a locked-down open wifi connection.

From TFA, this is apparently a huge improvement on previous attempts, but it's still not exactly dazzling. What sort of self-correcting protocol do you need to handle 499 of every 500 bits being lost?

If Mercedes has cracked the trick of 100% successful upgrades over air, great! If not, I'd prefer to know that the systems controlling almost everything on the Mercedes hurtling towards me is not going to die at some arbitrary moment. Bricked iPhones are inconvenient. Bricked 2-ton vehicles moving at 70mph are very inconvenient!

... the candidates could ask the person sitting next to them. Or maybe get a contractor to do the work for them. Or... at some point, exams have to stop being exactly like the real world. Otherwise, you need an exam that looks like and lasts as long as the career the candidate might take up on passing the exam.

Sorry for coming back to this old theme, but it seems to me that Mandriva just didn't have a niche to fill. Debian/RedHat have different parts of the server market. Suse has always had a niche market, and deals with other vendors might help. Ubuntu has pretty much cleaned up the "Just works on the desktop" market (whether or not that reputation is justified). Mandriva seemed to be aiming at something like the Ubuntu market, but there just wasn't enough room for two.

Yes, yes, yes to all the technically valid points about how distros are largely collections of files from elsewhere, about how any distro can therefore do anything if you hack it enough, etc. But, if we're talking about markets, what you get when you open the box is quite important. And what you seemed to get with Mandriva, at least at first glance, was "Like Ubuntu but with less resources behind it."

And yes to the possibility and maybe even the desirability of a thousand distros. But I can't see how more than a few of those distros can make serious money at any one time.

Exactly. A lot of the alleged technical commentary above is basically a worked example of how "free markets" do old-style protectionism.

Also, the computer revolution really took off because computers became ever cheaper and easier to manufacture. The problem with Babbage's design was that a lifetime wasn't long enough to build one without CAD/CAM. So, even if he'd worked twice as fast and got the thing working, it would have been a one-off for another few decades.

If Babbage had succeeded, it would have sparked the "man as machine" line of thought that has changed so much in our society, and that could have changed the course of history in all sorts of ways. Maybe we would have had wetware trials rather than monkey trials. But I don't see mechanical computing ever getting past the early mainframe stage in terms of numbers of machines or direct relevance to the average person's life. The Babbage Clone industry would really have struggled.

Apart from the observation above that this story is from the Daily Mail (which, among other things, campaigned against vaccinations for many years - hardly a pro-science position), there are lots of reasons to walk out of a lecture. In the case of Richard Dawkins, the man is intentionally offensive about all forms of religious belief, to a point that many atheists in Britain do not want to be associated with his rhetoric. Presenting evolution is one thing. Littering that presentation with, eg describing religious faith as if it is a psychiatric disorder. (Before everyone agrees with him on this point, could we remember that this was the pretext for the forced "treatment" of thousands of people in the former USSR, and that something similar happened to homosexuals in many countries. In other words, it's an argument that lays the foundations for some shocking human rights abuses.)

Walking out during a Richard Dawkins rant may, in some cases, be a statement *for* reasonable intellectual dialogue rather than tub-thumping atheist fundamentalism.

I live in France, where you can't get kids into kindergarten or school without vaccination certificates. And they cut child benefit too. The result is a very high vaccination rate, and that protects those who cannot be vaccinated such as very young children.

First, books are not all entirely sequential. One of the reasons I still buy paper books on IT is that I reckon I can often find what I want faster than by searching. Yesterday I wanted to remind myself how to iterate through a directory using Perl. I know the recipe is in "Perl Cookbook". I know that book is within arm's reach. I remember the chapter is about a quarter of the way through. Flick flick - bingo, all in about ten seconds. If I don't know the book well I look at the contents page, which is no slower than skimming links on a screen. Yes, I'm sure a computer is faster in theory, but that isn't my subjective experience, and I don't think that's because I'm incapable of using a computer. Grabbing a book and flicking to roughly the right place is actually not a bad random access heuristic.

Second, sequential is often good. It often *is* the pedagogy. When I first get a book (not a cookbook...) I often start on page one and read to the end, maybe skipping bits that really don't interest me. Yes, that takes longer if I'm looking for one specific answer. But, if the book is well-constructed, it often gives me a much better feel for the overall subject than I would get by looking at 200 modules, each of which is designed to stand on its own. And, in practice, I'd probably only read 20 of those 200 modules because there's no narrative to pull me onto the next module.

A great example of this for me is "Mastering Regular Expressions" by Friedl. You can google most of the answers to specific regex problems, and I did that a lot. Ploughing through hundreds of pages of dense and often obsessive text (breaking off from writing a chapter to get Richard Stallman to patch emacs regexes counts as obsessive, right?) meant that I finally understood how regexes work, what happens under the hood and why some apparently innocuous regexes never terminate.

Hypertext and modularity have their place. But I wouldn't dance on the grave of sequentially structured information just yet.