Real nerds who spend their days up to their elbows in Internet plumbing knew exactly what it was about at first glance.

IANAL either, but I know that passing the bar is a state-by-state test; you don't just take one test that covers the laws of all the states, you take a test for the state you are going to be licensed in (or the federal version, but that only covers federal law). So it's entirely reasonable he wouldn't know about the precise details of the law in "most states" even if he picked up a general knowledge in law school.

It's a mystery to me what "satisfactorily stable" consists of for people who point out availability as a problem with cloud solutions. As a rule, enterprises don't publish their internal downtime statistics, but I can tell you that for a large chunk of them, it's far worse than the occasional Gmail outages. And no one who makes that argument ever seems to look at the necessary companion to stability, which is cost. What does it cost you to be satisfactorily stable running internally? For most businesses, again, it's a lot more than an Apps subscription, for a lot less stability.

As for off-line access, Gears is already available to allow offline access to Gmail, but if you don't like that, you can just as easily configure the same sort of standard POP3 or IMAP client-side application that you would use with any other mail service, with the same capabilities should your connection be severed.

It also seems to me that people over-estimate what actually gets done in many offices when network access goes out, regardless of the off-line capabilities of clients, but that's another arguement.

Ironically, I think you have illustrated exactly why the desktop/laptop will be dead.

Similar arguments have been made for every outmoded technology by people who can't quite grasp that the world constantly changes around them. The only objection you are really leveling is already being attacked from both angles (areas with little or not Internet connectivity have been shrinking rapidly for a decade now, and many SaaS companies are providing basic caching and local operation of their services via browser plugins). We'll always still use something with some local processing capacity and storage to access cloud-based services, and in that sense, of course the "computer" is not dead... but it won't be the same processor/storage behemoth that we currently think of as a desktop or laptop. In fact, that's why they are calling those things you mention "netbooks"... because they aren't really a laptop or a desktop, no more than a horse and buggy is a car.

There are a load of fine suggestions in this thread which are well-constructed for logical minds, but I can't help but feel this tactic is best answered in kind: a gut-level fear-check. And so the best response isn't to sit down and try to explain the perils of security through obscurity, nor to try to sell additional security services, or to discuss patch cycles and the like, but instead to simply ask the client this: "When's the last time you heard on the evening news anything about a new virus, exploit, or vulnerability discovered in your Linux software? Now, how about Microsoft software?"

Overly simplistic? Absolutely. Sure to make them reconsider what the Microsoft vendors are trying to sell them on its supposed security? Definitely.

I think what it really shows the perils of is piling additional "features" on top of a perfectly good product until you've ruined what made it good in the first place and turned it into worthless crap. Search should be simple: give the user what they are looking for. All the other extraneous stuff they are loading it up with is bound to interfere with that basic requirement at some point.

I see this in mature development projects all the time. At some point, people get a pretty good product working, but they can't repress the urge to continue "improving" it... it can be boredom, wow factor for marketing, or just plain stupidity, but few people or organizations seem to know when to quit messing with a product that already works well.