If it's properly secure then open source the code. Security is in the algorithm, not the implementation.

Why not just use a one-time-pad? Get the launch crew to type it in on mission start?

Yes, this is a very different situation to Chernobyl but the worst case is actually far, far worse (in some ways of measuring at least). The problem here isn't the reactors themselves but with the spent fuel stockpile. Estimates have the potential for an uncontrolled meltdown in the spent fuel pile at orders of magnitude higher radiation exposure than were experienced from the Chernobyl incident, added to this exposure causing major problems in continuing to cool the reactor cores still under threat.

I have no idea what you're talking about "thousands of people to try to control the Russian plant" either. For a start, it was Ukrainian or Soviet but let's not stand on petty national boundaries too much. Second, about 30 people died as a direct result of the incident which makes it, uh, 0.03 thousand?

Not in my experience. Printers tend to be a crazy mess of different technologies supplied by the cheapest supplier/closest friend of the IT Manager/whatever someone found at Best Buy/etc. I'd estimate maybe 25% of the printers I've seen in corp environments support Postscript, about 50% support some variant of PCL (which mostly overlaps the Postscript ones) and the rest are a mix of custom drivers and just plain bizzare cruft.

If it makes you feel any better, the non-PS/PCL ones tend to not have x64 drivers for Windows so the whole thing just demonstrates the typical corporate shortsightedness in purchasing decisions.

Scanners in a corporate environment tend to be photocopiers with a network card that dumps a file somewhere so they likely won't be as much trouble as a printer. There's still the odd bizzaro scanner that just doesn't have drivers for anything but Win95 but those are slowly dying out. Assuming SANE is a very, very risky proposition.

Having a Windows Print Server doesn't really work because Windows works best by offloading the rendering to the client rather than using the server.

So, the print driver issue is likely real (though odd because it would have been cheaper to just get Linux compatible stuff for far less than driver development costs); the interop between OpenOffice and MS Office is definitely real; and there's more likely a lot of plain bad planning that just made a mess of the whole migration which put a bad name on tech that really isn't that bad.

That still doesn't explain why you think select() (or epoll) is actually better than IOCP. It more says "select isn't that bad", but says nothing about IOCP itself. I'm curious as to why you believe it is "deeply flawed" and "convoluted" to use completion ports.

Apple also sells their OS as a retail stand-alone item as well as the bundled item (go check the Apple Store). I think a valid argument could be made that Apple therefore is behaving the same way IBM is, but you would be hard pressed to define the Mac as a monopoly in any class of machine that isn't defined by the manufacturer.

How are you supposed to have a religious war if you bring some facts like that into it? Of course the API matters very little compared to the rest of the system, but when has that ever stopped Slashdotters going rabid over things?

Pick whatever API suits the platform you're targetting and run with it. Have to go cross-platform then use a different shim to the API of choice on that platform. Simple.

It's hardly a "random senator". It's the Government's Minister of Communications and he's pushing the fact that they've already done trials and found the filters "100% effective".

I don't think there's that much "widespread unpopularity" either - it's all couched in "if you're against this, then you're on the side of child pornographers", so people aren't really that opposed to it. It's being played very well by the people who want to control the net over here - never mind that once it actually goes in it will be a nightmare to control.

Seriously, what?

Sun has completely tanked in just about every aspect of their business. They're trading at around 1% of their peak value, and about 10% of their 5 year value. They're still laying off staff like nobody's business and they're really ripe for snapping up by some other company. McNealy drove them into the ground with a complete failure to read the market and respond to threats to Java and/or external influences on Java. Now you have the promise of Java revolutionizing the desktop all but dead to .NET, and IBM and Apache effectively in control of Java in the server space.

Taking McNealy's advice on technology is about as smart as taking Kenneth Lay's advice on energy independence.

If you really want some better advice, look towards some of the more successful companies in the software space and get a set of opinions to compare against each other. I'd probably take a range like IBM, Microsoft, Oracle, Apple for opinions and take into account their natural biases when you read their reports. McNealy's just wrong and so typical of governments rewarding failure with fat contracting positions.