"A fraction of the cost," while technically true, is far from the truism it used to be. Gone are the days where you could spend $800 and get the equivalent of a $3000 prebuilt system. In many cases, you're hard pressed just to break even today.

This is something of a dated way of thinking. In 2012, you don't usually see component failures; while it happens, it doesn't happen nearly as frequently as it used to. Therefore, "knowing what's in the box" is a value add of dubious value to many users. "Seagate, Western Digital, I don't care, I just don't want it to break on me."

Some people just want a computer, not a hobby.

Oh wow, it gets worse. Oracle won this with a $88.5 million bid; what the hell took the Air Force so long to pull the plug with that kind of overrun?

Seems that this is a common theme with ERP rollouts-- scope creep tends to get them all in the end. Granted, most organizations seem to wave off long before the $1 billion mark...

Well... yes. My employer runs three racks of servers all in; we don't have the bandwidth / R&D budget to investigate better options. The big players (Google, Amazon, etc) need to pioneer research in this area, at which point it will (ideally) trickle down to the masses.

Sometimes.

I'd argue this is part of the geek/hacker mindset, and while it's a valuable asset, we have to remember that this places us outside of the mass market in some fairly significant ways. As a direct result of this, we're no longer the "target market" for consumer electronics.

The problem with this approach is that it ties you to your ISP. When you move or they get bought in ten years, you have to try to recall EVERYONE who has your email address, and convince them to update their address books.

Postfix didn't exist until December of 1998. Before that, you were likely using Sendmail, or (god help you) qmail.

This was more sensible a decade ago; nowadays with so much of our lives online (banking, shopping, correspondance) it's no longer "reasonable" to not do anything "personal" on the internet while you're at work.

I very much agree with you. The entire *reason* that those companies have so many divisions is that they want goodwill you feel towards one of their products (say, the Discman of old) to transfer over to other lines of products (say, their headphones).

However, what this means is that when one division (or, in this case, several) radically screw the pooch, a lot of people associate the negative experience with the company as a whole. Ergo, due to the CD / DRM issue almost a decade ago, I won't buy a Playstation, a VAIO, or a $20 pair of headphones that say Sony on them.

It's not just *GOOD* feelings that transfer over, Sony.

Yeah, not trying to be offensive here, but answering the questions you've posed has spun up an entire industry; it's decidedly non-trivial. On the plus side, for a project of this size you can quite easily get a number of consultancies in Chicago to quote you free of charge.

Assume a DC-wide power outage. You power things back on.

If you can't come back up without the controller (which is on a VM) you fail.

Now, I'm not a VMware guy, so you'll have to tell me-- what does this failure case look like?

Or tor. Or VPN endpoints overseas. Or ssh tunnels.

I don't really see how legislation can reasonably expect to keep up with technological innovation.