I agree that Mao did help to drag China into the modern age (kicking and screaming albeit) but it was Xiaoping after Mao's death and Zhou Enlai beforehand who really set China up as a superpower.

In my opinion Xiaoping and Enlai achieved more amazing feats (against biao and jiang qing and the others in the gang of four clique no less) than Mao did.

Mao also made gigantic economic mistakes such as the great leap forward which led to the deaths of millions of chinese via starvation due to economic mismanagement and an irrational desire to buy nuclear weapons at huge costs to the chinese people.

Still, Mao's role in the original revolution, removing the corrupt kuomintang and unifying, modernising china should not be discounted.

Alternatively, simply be so *awesome* that you escape capture and destroy the entire universe.

Firstly, debian already ported most of Linux software to ARM, software availability is not an issue anyway.

Secondly, x86_64 is an extension on x86. Linux32 is just a set of 32-bit libraries compiled against a 64-bit kernel, that allows you to run 32-bit apps, using features of the processor specifically designed to do this. ARM is a completely different architecture and such an approach is simply impossible. The only way to run other x86 applications on ARM are via virtualization, which frankly would be unusably slow on a netbook.

Exactly. Debian already distributes an ARM version of their entire distribution.

Debian has a complete ARM distribution including all of those things you describe. It wouldn't be hard for Ubuntu to shift their distribution efforts to ARM. In fact, it's just changing a few lines in a shell script.

The porting to Qt isn't really necessary, recent versions of chrome have heavily abstracted the UI so switching to a different platform is a matter of changing a compile flag.

Microsoft struck a deal with a Japanese software company and ensured that Windows had perfectly localized versions available in the Japanese market. It took off rapidly and the Japanese market is still more MS dominated than most of the rest of the world.

Europe did have numerous home-grown systems but they were all gradually eclipsed by the cheap and extensible IBM PC which naturally had MS-DOS so Microsoft got a foothold.

The program is formally proven to be correct, where "correct" is defined to mean conforms exactly to the specification.

The specification may not be useful, but it does exactly what the specification says it does.

If I have a detailed specification for an entire operating environment (which this is not) and it is all formally proven to conform to said specification, then I would happily use it over something that had been used alot and didn't crash often. If it doesn't do what I want, then that means my specification was inaccurate.

I think such an application is perfectly justified, but then again I teach postgraduate students how to write haskell in haskell.

In a hypervisor system, not even sure if that's possible.

It doesn't return anything about success or failure.

Yes.

Do you know what a microkernel is, and why they are so forcefully advocated?

I work at NICTA from time to time. Will raise this with them.

Haskell's garbage collection works (or should work) based on formal inference -- seeing as they translated the haskell to C manually, they would have proved the garbage collection as part of the deal. Also, haskell's prelude and basic libraries have already been formally proven.

Because of formal inference, a good Haskell compiler knows exactly when and what will be garbage collected at compile time - there is no mark-and-sweep approach like Java. It's no worse that manually freeing stuff in C, in fact it's substantially better.

L4 doesn't delegate memory management to a process, that is impossible, because it would then have to manage memory for the memory management process.