That's not completely true. Millions of people around the world paying small amounts of money for iOS apps shows that it can be done. Additionally, I signed up for a $1.29/month charge for an extra 20G of iCloud storage, and it was extremely easy.

If Apple can figure it out, why can't spotify? Or is the 'enter your credit card' details step the part that you indirectly refer to as difficult? If so, then sure, there's no easy way of charging people a small amount of money without requiring and storing credit card details. If you can get people's credit card details, then what's the problem?

Of course, once you've launched a free service - and rightly or wrongly people really do expect free stuff - it's rather hard to convert that into a 'pay' service without losing a large percentage of your user base.

And of those scripts you link to, two stand out as being simple and to-the-point. The others are dreadfully complex, and *different* for every flavor of Linux. Utter madness.

The Systemd and Launchd ones are declarative statements to the system about how to handle the daemon. Seriously, what is not to like?

Why people would be in love with bash scripts and PID files to manage daemons on their OS I continue to be baffled by. In the meantime, I just use systemd and move on with my life.

Legal. Displacing. Dying.

Anyway. What does this story have to do with immigration specifically? Would the situation be considered acceptable if the IT department were laid off, and lower paid Americans were contracted in to continue the work? And if so, why?

This may be true of spoken language, which humans develop naturally and as a matter of instinct. But it is manifestly not true of the written language, and suggesting that your inability (or unwillingness) to correctly capitalise and punctuate simple sentences is not a form of rebellion against those who would control you.

Forms of control exercised through control of expression are, of course, very real. And were I subject to any of them I could imagine that they would also be very effective. But having grammar rules that aim to ease written expression and comprehension is not a form of control in this sense.

The rules are arbitrary, up to a point, and they evolve over time too. But they don't evolve very fast, unlike spoken language, and it is because of this that I am still able to pick up a book written a hundred years ago and read it without too much trouble. And in any case, the article you link to is not a screed aimed at dismantling grammar, rather it suggests that while grammatical rules do exist, they are not as described in the Elements of Style.

This is a tautology. Also, that level of complexity might be quite low.

Means that there are two people in the world who match the DNA evidence. If the DNA evidence is the only evidence, then the chances that you've got the right person is 50/50.

DNA evidence alone should never convict, only exonerate.

Pathetic fallacy. Please rewrite.

Either you're a grandmaster, or you're a long way away from achieving this. I assume that it's the latter, but please forgive me if it's the former.

Sorry, I forgot to address this. I'm not suggesting multiple checkouts to avoid branching, I'm suggesting multiple checkouts when working on multiple branches to avoid rebuilding when you switch from one to another.

Rubbish. In SVN you can branch, and switch your working copy to the branch in a single command-line call. With the advantage that the branch exists in the central server (yes, this is an advantage) right away so other people you work with can work on it too.

The cost of branching is not related to how easy the tool makes it to create a branch. Contrary to common belief, it's equally easy in both SVN and git to create a branch. The cost of branching is equal to the cost of merging those branches together. Git is no magic bullet here, if you have 'n' branches, then the cost of merging is O(n^2-n).

Then I will check out a new copy - disk space is cheap and my time wasted waiting for a build when I switch back again is expensive.

If I was working out my last couple of weeks, and IT put their fileservers on read-only for me just in case I decided to act like a dick, I would be deeply offended. And I would walk out of the building and not return.

Well that all seems pretty above board - I wonder then, given how hard it often is to retrieve the black boxes - why commercial passenger planes don't stream at least some of their flight data back to a central location somewhere?

Yes of course they do.

The Arduino tools are appalling, literally the worst development environment I have ever used. And I've written code in emacs. I do appreciate that many of the competing products are a little bit too complicated, but that Arduino thing is a shocker. Surely there's a middle ground somewhere?

The forums are... well.. not very helpful.

To be fair though, the library support that ships inside that dreadful "IDE" is quite good.

You could do that with all the other ones too. TI's Launchpad, Freescale's KINETIS board, STM have their discovery boards. They all let you blink some LEDs in a matter of minutes - and the Arduino ones are the most expensive. I rather suspect that most people don't do much more than blink some LEDs anyway, since doing anything much more complicated than that with the Arduino "IDE" is an extremely painful exercise indeed.