What's the problem? The NSA can always get your data back for you? Right?

Lots of job advancement opportunities for Number 3 thru N guys at Al Queda. The trick looks to be is to find another gig before you advance to Number 2..

As for Mr Oerling, I think he is probably delusional and is vastly underestimating both the number of serious security flaws in modern software and the number of folks attempting to find and exploit the flaws, but maybe he knows something I don't.

Almost everything electronic has been replaced with higher complexity, yet still higher reliability,

Half right. Stuff does tend to become more complex over time. But not necessarily more reliable or more usable. For example, some of our kitchen appliances are indeed more usable than those I grew up with in the 1950s. Some are cluttered with unnecessary, weird, or incomprehensible "features". There are a couple of companies whose products I won't even consider any more when making purchasing decisions because of the loathsome controls they have inflicted on the world in the past.

Reliability? Thanks largely to the Japanese -- who actually care about such things -- Automobiles actually have become more reliable over time. Kitchen appliances. Not so much. Electronics? Look no further than the shambles that the internet has become.

Complexity is not necessarily good. I suspect that if the newer is better school had their way, hammers would weight 20 kilos, have an incomprehensible control panel, 17 moving parts, three circuit boards and would require both power and internet connections. And you'd need to log into them

So it seems to be less about the medium and more about the designed controls.

Exactly. If designers want to do internals with digital bits, that's their decision. If they even have a decision. But the output should be adjusted or adjustable to user needs. Which is hard, because frankly we techies suck at interface design and .experts on interface design seem to be, if anything, worse than non-experts at producing usable devices. For situations like trying to adjust for maximum or minimal level, digital readouts can be pretty much unusable and it's hard to beat the classic analog design that Wikipedia tells me has been around for two centuries.

However, I should probably add that using a web oriented computer like Chromebook in a place with only dialup might encounter a few problems.

All I know about Chromebooks is what I just read at Wikipedia. But they seem to have USB ports. There certainly are USB fax modems out there and they are cheap. I just set one up on a netbook so we can ditch the ancient fax machine pushed back in a corner of the bedroom. I don't know for sure that the Trendnet USB modem we bought works with Linux/ChromeOS, but it certainly runs well enough with Windows XP to get a login prompt from a local Netzero node.

My proposal is that airlines add a new seating class to be known as Midget Class (colloquially "Sardine Class"). MC will be available only to passengers under 160cm (5ft 3 in for Americans) and 50kg(110lb). It will be priced the same as Economy and Economy will be redesignated as DeLuxe and priced at 1.4 times MC. MC seats will be smaller and stacked vertically and horizontally using a sophisticated packing algorithm. The legroom in Deluxe will be reduced by 2.5cm (one inch) from the current Economy.

BTW, between the steadily shrinking seats, nutty security theater, inability to maintain published schedules and third world chaos of airport operations, I quit flying a decade ago. I realize that not everyone has that option. But I would ask those that do, why they pay money to be subjected to modern airtravel? Buses, trains (even US trains), cars, and ships are far less aggravation. I'm far from sure that Hitchhiking isn't about as comfortable and reliable.

Maybe you're both correct to some extent. Rebooting certainly doesn't solve all problems. But the software architecture used in Windows/Unix does have the unfortunate characteristic that it sometimes manages to transition into states that no one anticipated and that do undesirable things. Rebooting restores a more desirable state. At least for a while.

There is also a problem that few modern PCs use memory capable of detecting memory errors Thus it's possible for values defining system state to change spontaneously without being detected. That's an interesting case of shifting costs from visible hardware costs to less visible support costs -- largely Microsoft's (bad) idea BTW. Long story there. Anyway rebooting will help if important bits somewhere in memory have reset themselves.

There is some credible evidence that flaky PC memory is more common than most people assume. See http://en.wikipedia.org/wiki/R...

OTOH, if the problem is a logic error in code, or bad documentation, or an atrocious user interface, or the user -- rebooting can't fix it.

Translation: "Why should I learn from the mistakes made in the past? I'll just make them all over again.

I doubt you'll manage to repeat ALL of the mistakes of the past. You'd have to be pretty clever to do that. But with sufficient effort and diligence, you can probably manage to repeat most of them. And maybe even come up with one or two truly innovative cock-ups that we old timers overlooked.

Personally, I'm not fond of replacing real world testing completely with simulations.

Exactly. A broad battery of simulations makes sense for regression testing to prove that the 2027 model year software handles all the situations that the 2026 did. But real world testing is required to verify that the system doesn't do nutty things when confronted with unusual conditions -- dust clouds, ice coated wall to wall potholes, a trackless rural road or rarely used off ramp covered with four inches of snow with whiteout conditions (where is the edge of this damn road?)..

Keep in mind that we treat human misjudgments in extreme conditions as unavoidable accidents. But unless I misjudge Homo lawyerensis (a species that regretably perhaps does not appear to be endangered), every significant accident involving a self-driving vehicle is probably going to be the manufacturer's fault.

MSDOS certainly was simple: it was 16-bit, it lacked preemptive multitasking, and each program was limited to 64kB of memory (that other processes were not prevented from overwriting)!

MSDOS also worked perfectly adequately as the centerpiece of Windows 95 and 98. Of course working is important only to users. And who gives a damn about THEM? (Unless they get fed up enough to leave)

Oddly enough Microsoft's stock price stopped rising about the time that NT started to replace Windows 9. And the rather widespread dislike of Microsoft started about that time.

Just coincidence, I'm sure

BTW, the overly complex OS (relative to current low end device capabilities) is only one of the problems MS faces. And probably not the largest. The multitude of poorly documented and idiosyncratic APIs is probably a bigger issue since the principle reason for selecting Windows is likely in many cases to be compatibility with old applications. If Windows 2016 can't support the software some dude wrote for you in 1996 to control your packaging machinery or deal with your peculiar audit requirements, you probably aren't going to buy Windows 2016.

Having about computers BEFORE there was such a thing as computer science, I have to confess, that I have never been able to appreciate the vast benefits of patterning operating systems on those used by 1970s mainframes. Truth be told, they often didn't actually work all that well in the 1970s. Sometimes still don't if you ask me.

Some "advances" in computing really have made things better/easier. Higher level languages with exception handing? Absolutely. File systems vs rigidly allocated mass storage? Terrific. NT vs MSDOS? No Hum. Not in the same class with meaningful advances in computing.

All other things being equal, I would probably go with NT. But all other things don't seem to be so equal. MSDOS was simple and ran well on minimal hardware. NT isn't simple and doesn't seem to run all that well on slow CPUs. We have a couple of EEE PCs around the house running XT and Windows 7. They are both terminally slow. That's not entirely an OS architecture issue I think. Unix often seems to do much better on lightweight hardware (as long as you aren't trying to print). But the NT architecture probably doesn't help and it's always been unclear to me exactly what NT brings to the party on a lightweight personal computer -- which is, after all, what all those itsy devices whose marketplace Microsoft is having trouble selling into are.

Now that hardware has advanced they have a much better shot at architectural unification

Trouble is that the hardware has been marching off in directions often orthagonal (or worse) to the direction software and applications have taken. Hardware now includes many slow, limited CPUs that allow reasonable battery life in very small, compact devices. Software OTOH is written, as much as is possible, with no concern whatsoever for resource usage. The result is huge, impossibly complex (and therefore not very secure), OSes with often agonizingly slow UIs.

I personally doubt anybody or any group of anybodies is/are clever enough to "fix"/unify Windows.

Maybe if Microsoft had made different decisions in the mid-1990s when they had a compact real mode OS with a usable GUI running atop it, they could have ended up with something unified or unifiable. But that was then and this is now and the intervening two decades are water under the bridge or over the dam or something.

I'd like to be wrong about this, but I doubt I am.

My intuition says that most people do this. Though, I could be wrong.

Well, some of us try to do it. We are, regrettably, impeded by whacked out sysadmins who insist we must use THEIR idea of a strong password -- which always seems to be different from anyone else's idea of a strong password, and/or that we need to change passwords periodically, and/or that we can't reuse passwords.

I sometimes seems that there is an inverse relationship between the actual need for security and the system administrator's perception of the need for security.

But other than the fact that users often have to contend with the idosyncracies of sociopaths who feel that anything that is easy to use is clearly flawed, this seems a pretty good idea. If it gets the attention it deserves, perhaps it might be one small first step toward straightening out the incredible mess that is computer security.

Actually, as a climate skeptic, I've been saying for years that we should all focus on innovative nuclear technologies.

In fairness, some true believers in catastrophic warming warming do support nuclear. In particular NASA's James Hansen -- whatever one may think of his analytic skills -- is an outspoken supporter of replacing fossil fuels with nuclear. However we do need to keep in mind that even a well designed nuclear plant is likely to be managed at times by incompetents -- political appointees, fools, risk takers, or the just plain crazed.. We need nuclear power plant designs that even TEPCO couldn't turn into a regional or global disaster. While such designs are conceivable -- e.g. pebble beds -- they do not currently exist in proven form. And without fail safe designs, large areas of the planet are -- and probably should be -- pretty much off limits to nuclear power.

Is that a solvable problem? Probably. Is anyone trying very hard to solve it? Not that I can see.