The problem is that it's already a crowded market. And furthermore, Amazon, Google and Microsoft have all discovered it, and now want their slice - and they all already have solid cloud compute platforms to use as a backend; so it's going to get even more crowded in short order.

(Full disclosure: I'm on the MS team that is working on the Azure IPython/Jupyter notebook service that just went live on PyData, which is one piece of that.)

If you get software and don't pay for it, guess what- it's free. To you.

Microsoft "shared source" licenses were actually more restrictive than that, as they had only allowed read-only access to the code, even for private and personal use without redistribution - i.e. fixing a bug and building your own version would already be a violation of the license.

If that is really what is meant by "accessing the source code" by the article submitter, then this is the way to go. You can just use MS-RSL verbatim for that. Just, please, don't call it "open source". Even Microsoft, back in the day when it was still doing this kind of thing and calling GPL a "virus", had the decency to not appropriate the term and use "shared source" instead.

And note how this has largely failed for Microsoft. Very few people actually bothered to do anything with those "shared source" projects, and there was no community to speak of as a consequence (and as you rightly note, it's a misnomer anyway). I don't think there are still any MS projects shipping under any of these licenses anymore; pretty much everything is either AL 2.0 or MIT now, including things like CLR that used to be "shared source".

Thing is, given that Python and R have set the bar to F/OSS effectively (at least for the "core" that's necessary to run things, if not the associated tooling), anything less than that will quite rightly be seen as deficient. Whether they have something else to offer that could compensate is another question, but from my experience talking to people in that community, it's a fairly major point that I doubt they can easily compensate for.

For the scientific community, this has actually been more important recently as people want to be able to reproduce other people's experiments. Which is part of the reason why why F/OSS tools like scipy and IPython/Jupyter have been growing very fast.

There is a very big difference between just dumping things into the oceans as is (which is what the article you've linked to is about), and encasing them into a strong protective shell that prevents leaking, and picking specific places that are studied in advance to guarantee enduring safety and proper long-term disposal.

There are some valid objections to using subduction zones for this, but everything that I've read indicates that it is, at worst, an engineering problem that could be solved if desired, not a dead end. The only reason why it's not seriously explored is because of the treaty prohibiting it, and the treaty was originally intended to deal with exactly the kind of thing you've linked to.

To see the actual minimums, look at the lowest-end devices that are sold with 8.1 preinstalled. These would be the 8" Atom tablets, like Asus VivoTab Note. Most of those are 1.6-8 GHz CPU and 2Gb RAM. That's the lowest end of the "comfort zone".

Remember that 8 (well, RT, but it's basically slightly trimmed 8) had to run on ARM tablets. So 7 -> 8 also had a lot of optimizations done for sheer necessity of it.

Political problems are self-induced. When there's a genuine engineering or scientific challenge to be overcome, we have an excuse for not doing it just yet. But when all that stops us is some form of NIMBY or "la la la it's not happening" or "I just don't care", there's no excuse, just stupidity. Same goes for AGW denialists, anti-vax etc.

The standard C++ library for multithreading is <thread> - <threads.h> is a C thing (which is still available in C++, but not preferred).

None of this make much sense. For starters,

The semantics of batch files is that every statement runs sequentially. This cannot be changed because all existing batch files assume that the previous line has completed processing, and the next one can use its results. Consequently, it doesn't make any sense to "multi-thread" a batch file, since it's entirely sequential.

Additionally, batch files spawn processes, not threads, so "multithreaded" is a misnomer here in any case.

Okay, now you really don't know what you're talking about. First of all, .NET is not really tied to Windows, and it has the same behavior regardless of which version it runs on. Second, GC always runs on a separate thread, and this has been the case since 1.0. Third, WinForms is a Win32 wrapper, and as such it has to run on the main thread, which is trivially observable in debugger - and nothing has changed here in Win7 or any recent .NET version.

Yes, exactly. We have been banning obvious solutions to problems, and then complaining that nuclear is hard. It's ridiculous.

You forgot natural gas. Though that may have unfortunate geopolitical implications depending on where it is.

Why would you want to discharge tritium? It has many practical applications (and is fairly expensive because of that).