Also, server is not their cash cow. Between desktop Windows, consumer hardware (including the Surface, mice, keyboards, etc) and business software they pull in about $44B. Their server products, including Azure, MS SQL, consulting services, and development tools only bring in about $20B. Entertainment and mobile bring in another $10B.

Microsoft benefits from having a broad, comprehensive product line, whether the divisions themselves are necessarily profitable. Just like Apple, Google, and Amazon, they benefit from getting people to buy into their ecosystem.

They have struggled in recent years, not just because Apple's resurgence has eroded their dominance in the PC market, but because the market has shifted away from the PC model. More people are using tablets or even phones as their primary device, which means more people using iOS, Android or Amazon devices, and their associated app, content and services ecosystems.

Their response seems to be increased openness (see: open sourced .NET releases, patent pledges), interoperability (see: JUniversal, AllJoyn contributions), convergence (see: Windows 10, from mobile to server), and free stuff (see: Windows for small devices, hobbyist platforms; Visual Studio Community Edition etc). I find it hard to fault that.

Lumia ring a bell? Windows 10 is their convergence release, with mobile being just one more edition. Plenty of phones with far less horsepower are running Windows. Plenty of tablets with similar horsepower are running Windows.

I picked up a couple of Stream 7's, which also sport only 1GB of memory, a modest low power processor and a full copy of Windows - as stocking stuffers this year. They work fine for what they are. Especially since I paid $74 for one and $49 for the other, thanks to in-store "shopping events".

This strikes me more about the hobbyist community and pervasive computer than the "internet of things". This board is not suitable for those applications, regardless of what platform it is running. But this has potential for more demanding embedded applications, internet kiosks, network terminals, or even starter computers for kids (which is what I'm considering getting a Pi 2 for).

I probably won't land on Windows 10 as the final operating system; I haven't used Windows as a desktop in about twenty years now. But I'll certainly play with it.

Whatever you think of his ethics or achievements, it's hard to argue that Bill Gates was not (or is not) a nerd.

Yes, yes, TANSTAAFL.

But as many free software advocates point out, copying software IS as close to free as it gets. There is no opportunity cost to Microsoft, since the chance that people would otherwise pay for a copy of Windows to run on a Pi is essentially nil. There is little opportunity cost to the user, since they would otherwise be installing and configuring an operating system anyway. Even if they ultimately choose not to keep it, the time and effort involved is fairly negligible and there is arguably value in the experience. (At least to that individual, since they freely chose to engage in it.)

Really? If an afternoon trying out a new platform represents a significant investment to you, your time is far more valuable than mine. Most people I know spend more time dicking around with the lastest MMORG on a regular basis.

As for programming, Windows remains a highly lucrative development platform, regardless of your personal opinion. Even in the mobile space, targeting a platform without a saturated ecosystem may ultimately yield better results. Easier for an app to stand out if it is one of a dozen instead of one of a thousand.

You assume that maintaining another ARM port takes an appreciable number of developer hours, that code quality isn't generally improved by maintaining ports on disparate architectures, that community goodwill has no value, that community testing and input has no value, that this is an "unsuitable platform" in the first place (the Pi 2 is more powerful than a lot of phones they run on, for example), that it won't net any new customers from people who are curious to see how Windows 10 is, etc.

As initiatives go, this is doubtless cheap and has plenty of positive externalities. And from a user perspective, trying it out is a throwaway afternoon at worst.

So installing a free copy of an operating system will make you lose money or give you a disease? Didn't think so. Grow up.

Know what systemd was heavily inspired by? The MacOS X init, launchd.

You can find a full smart phone for $30 in the United States as well. But not everyone wants or needs a full smartphone (see: parents, elderly). Or they care more about battery life than features.

No one?

ahref=https://twitter.com/jruby/status/534471839620165632rel=url2html-15138https://twitter.com/jruby/stat...>

Considering the proliferation of programming languages out there, holding a spot in the top ten is respectable. Certainly not tantamount to self-mutilation.

It may not be the most popular option, but Ruby is hardly a marginal language. RedMonk has it tied for 6th with C++, PYPL has it at 10th, and TIOBE has it at 14th. It came off from the Rails high, but it remains steadily popular.

The ecosystem has actually got significantly better over the years, especially as Puppet, Chef, MCollective and others have driven popularity as an admin language, rather than a web language. But more importantly, JRuby pulls in the entire Java ecosystem, which actually puts it in a better position than perl or python, in my opinion. There is Jython, but that lags significantly behind C Python (current stable is 2.5 compatible, which was released eight years ago; their 2.7 release has been in beta for about 21 months) while JRuby offers Ruby 2.1 compatibility in their current stable release and will be putting out their release candidate for 2.2 around the same time as the Ruby 2.2 stable release.

The real win for me is JRuby. The Java ecosystem is at least as broad as perl, and generally better suited to enterprise applications. There are generally perl modules for everything, but they often perform far worse (e.g. Net::LDAP is probably an order of magnitude slower than UnboundID processing LDIF) or are just terrible code (e.g. Net::Sieve::Script which is a regex-based hack, rather than an actual language parser like jSieve).

1) anyObject.class

2) Um, no. Objects cannot magically change their class, period. You might assign a different object to a given variable, but the language is strongly typed.

3) Huh? There are relatively few symbols in Ruby, as a rule. Are you referring to special variables (like $: $0 etc?)? Those are ancillary and not considered idiomatic these days. Don't like them? Don't use them.

4) Why shouldn't they? The first is a just a chained method. No different than "foo.split(' ').length;" in Java. I'm assuming the latter is supposed to be "num.to_s 16", which would be "Integer.toString(num,16)" in Java, but that is just because the Java designers weren't nice enough to allow you to pass a radix argument to the non-static method. There is nothing in the language that would have precluded "num.toString(16);" being valid.

Java is grand. Using jRuby to prototype, explore and extend is even better.