The video is perfect. The video correctly represent Windows 7 and its design philosophy. And it evokes the same feelings in me that using Windows does.

Congratulations!

Well, if Microsoft's new OS can handle multi-core, multi-processor transparently for the applications

No more than current OS'es. This OS simply claims to be internally more efficient.

The thing I found quite elegant in Erlang is that it makes it so transparent

Erlang really does little that you can't do as easily in other languages. The real value of Erlang is in what it lacks: it prevents you from doing things that are hard to distribute across cores.

Imagine an OS with a "normal-looking" set of library that can handle all the hard works transparently. I'd say, bully to them.

That's wishful thinking. "Normal-looking" code is "normal-looking" because it uses constructs that are intrinsically hard to parallelize.

Microsoft is getting more like the old Xerox and IBM every day.

Not really. Xerox PARC did create entirely new technology. Xerox also turned that technology into products-good, if expensive, products. The technology didn't catch on because Apple undercut them with a low-cost knock-off.

There's little "industry changing" about what Microsoft has been releasing: microkernels, message passing, database technology, sandboxing, kernels in functional languages, that's all old tech; Microsoft is just implementing their own version of it.

The only reason people are paying attention to it is because of the power Microsoft holds in the market; if they finally update their totally obsolete Windows platform, that's news, even if they merely update it to something slightly less obsolete.

"Marshalling" means converting data structures into byte streams. No, you didn't have to do that multiple times. The term you're looking for is "routing". Routing can be abstracted into libraries and the OS; no need for every application to worry about it. It was just that the Transputer (as well as a lot of other system software development) was killed when Microsoft monopolized the market.

They do have a connection: politicians do the opposite of what they say.

Like when the "family values" politicians do prostitutes and drugs.

I am generally talking about PC's, here- and I can't think of what these IBM systems would offer that market.

What does the size of the hardware have to do with innovation? The fact remains that none of what you claim as "innovation" is actually new.

but technology-cleansing and losing the server market was all Unix.

By the time Microsoft had its opportunity with NT, UNIX was already legacy technology. UNIX would have been dead if Microsoft actually had actually come up with something better. And, believe me, people like me gave it a serious try.

It was Microsoft who missed their opportunity to innovate in operating systems, by betting on Cutler and NT and delivering a warmed over VMS.

leaving a VMS-based system like NT in a competitive position. Systems like NT and Unix are cognitive, they interface well with people. They won in a human market, not a technical market.

Yeah, that was a "human market" alright. A "human market" that consisted of Microsoft pressuring its business partners into shipping NT through illegal tying and bundling deals. Without Microsoft's marketing and business muscle, NT would have been DOA.

People with a base and some equipment (microscopes, spectroscopes, chemistry gear) would be extremely useful.

A dozen astronauts vs. one unmanned probe? Perhaps. (Of course, in several manned touch downs and sample returns, we didn't find any solid evidence of water.)

A dozen astronauts vs. 10000 unmanned probes? The probes win hands down in terms of utility, longevity, and scientific output.

And if you look at the costs involved, it's the dozen astronauts vs. 10000 space probes that we need to consider.

That's completely and utterly unrealistic. If those techniques and languages produced workable products, some would exist.

Since the 1960's, people have produced numerous kernels in languages other than C. Systems had virtualization, elaborate access controls and security architectures, and tons of other features. These weren't academic exercises, they were the workhorses of the computer industry. IBM alone had a highly virtualized product line, plus an entirely separate product line based on virtual machines. People were working on highly parallel systems, multicore support, verification, sandboxing, kernels in functional and managed languages, and all sorts of other "modern" things.

All that came to a halt in the 1990's. It came to a halt because Microsoft took over the industry, not with better technology, but with its illegal monopolistic practices. After Microsoft's takeover, neither academics nor companies felt there was much point in working on kernels anymore because Microsoft would not let anything succeed in the market anyway. It's going to take us another decade or so to recover from the Microsoft dark ages and for people to start working on operating system research again.

It sounds like you have a very absurdly ivory tower perspective anyway.

And you sound like the typical guy who was hired into Microsoft fresh out of college and has never seen anything else.

As for your "multi-role" arguments, you really need to read up a bit on the history of operating systems since you obviously have no idea of the history of this.

That's why we need manned space exploration: if we only had sent people instead of probes, they would have found this long ago!

Oh, wait...

I have never owned a system where Linux "just works". Most users purchase computers with licensed and customized operating system images.

And if you purchase your computer with Linux pre-installed by a vendor, it will work even better.

Replacing these with a hacked together 1970's operating system designed to work over dumb terminals is simply beyond retarded

Not as retarded as keeping an overpriced, hacked together 1970's operating system from a bunch of people who think that "multi-role" support is the future and that Amiga was the state of the art in operating systems in the 1980's.

Linux on the desktop is a non-starter. Its marketshare reflects this.

It is, although that has mostly to do with Microsoft's monopolistic practices. Fortunately, "the desktop" itself is going away. Hopefully, along with it, the incompetence up in Redmond will go out of business as well and the industry can finally move beyond both Linux and NT.

In terms of total number of installed Linux kernels in the world, Linux probably already has NT beat hands down.

What more modern approaches?

Managed programming languages, virtual machines, etc. All of that was around in the 1980's already.

Can you name a better language for writing kernels? C is basically the optimal language for writing system code. You haven't presented a revolutionary alternative yet.

Modula-3 and CEDAR/MESA come to mind, but kernels have been written in managed languages.

And C is about as far from being "optimal" for writing a many million line kernel as any language can be.

Virualization is big because UNIX is too awkward to work as a multi-role server

UNIX and Linux actually offer all the "multi-role" support you could possibly imagine and want: access control, isolation, namespace manipulations, and various forms of virtualization. Among all of those, people choose virtualization because it's the least hassle and the easiest to manage.

And you are absolutely right that NT has focused on "multi-role" support, including elaborate security and user mechanisms from the start. But that's not the sign of a modern kernel, it's a sign of an obsolete 1960's design. By doing so, Microsoft missed the boat on all the other approaches and is now trying to catch up.

Let's just be clear about this: Linux gives users a choice about what level of "multi-role" support they want, and their choice is "very little".

so you use virtualized myopic UNIX servers to replace what could otherwise be done by Java EE or .NET.

That claim makes absolutely no sense at all, since the Java EE approach was developed on UNIX and Linux. And once you run your servers in managed virtual machines, you don't need all the elaborate kernel-based "multi-role" support anyway. That's another reason the "multi-role" support in NT is superfluous and obsolete.

The only other "modern" systems I can think of from the beginning of the NT era are Amiga and eventually BeOS.

Yeah, that's because you obviously don't know anything, and neither did the people at Microsoft. Microsoft's OS developers in the 1980's and 1990's were a bunch of PC hackers plus industry wash-ups who had no idea what the state of the art in computer science actually was, and they developed a third rate OS that was obsolete from the start.

Let me tell you; its because neither you nor anybody you care about are affected.

You say that as if it's a bad thing.

Let me give you an example; between 1998 and 2004 about 4 million were killed in the second congo war; and that was a fairly localised conflict.

And in what way are we responsible for that, or responsible for fixing their problems? The most likely outcome of interfering in their wars is to make things worse, not better.

A basic problem with both of those packages is that they're database applications without a database.

It's a lot of work, but if you want high performance and/or have limited compute resources, that is often the way to go.

The big commercial DNS systems are all database-based.

They probably also throw 10x the amount of hardware and personnel at the problem than is strictly speaking needed.

Does the word "cloud" have any particular meaning?

I always thought it referred to a visible, translucent collection of tiny water droplets suspended in air high above ground level.

I don't have to supply facts, you do: you claim that Windows is objectively better than UNIX on a number of dimensions (e.g., driver development time, bloat, energy efficiency, etc.), but you don't supply a shred of evidence. Therefore, your claims of Windows' superiority is just a bunch of hot air (and quite implausible, too, given how Linux eats Windows' lunch in the embedded market).