> You can't "prove" that anything but a trivial program works correctly.

No, that is not at all what he proved. He proved that unless you limit yourself to trivial programs, there are always going to programs where you cannot reasonably figure out whether they work correctly. It is the mathematical proof that bad (as in difficult to understand) programs exist.

That is directly relevant to this discussion, but only the incompetent will use it as an excuse for delivering bad programs. All you have to do is avoid writing programs that are difficult to understand, and you are fine. That is one of the key rules of writing secure programs: keep it easy, keep it simple, and do peer reviews.

> You want bespoke (nuclear powerplant-level) security

No, you misunderstand the point completely. This is not so much about functional guarantees of the software than non-functional ones. If you buy a game for $0.99, you do not expect it to be bug free. However, you do (reasonable) expect it not to steal your credit card data or compromise your system.

It is just as with any other item for sale: a cheap toy made in China may not last very long, but you do expect it not to harm anyone. Every toy sold for $0.99 carries this guarantee, and the manufacturer is liable if anything happens. So why does software not?

Security is typically a non-functional requirement. And in a lot of software that is actually a reasonably easy problem. Sandboxing, restricting permissions, limiting network communication, and with just a few easy measures you have massively improved safety and security. That is what best practice means: what is known to work and reasonable. There is no need to assume that your code is bug free - power plant level security and testing is not required.

Yes - in '95 the internet connection of my university was over ISDN. Slackware was available locally on the ftp server. Those were exciting days.

> Firefox fixed that problem ages ago.

No, unfortunately not. They applied some kind of ugly clutch, but not a proper fix (such as a stable properly versioned API)...

> I think any corporation that picks Android would only give their employees on or two models to pick from

In fact as a corporation, you actually have little choice but to specify the Nexus range (or rather model). Nothing to chose from, but you know that you get at least a minimum of support. No other Android model currently offers that promise - and if you ask me, that is the problem of Android in a nut shell.

Don't get me wrong, the Galaxy Nexus is great device, but it is no Galaxy SIII, and it is certainly not a "one size fits all" device.

Haha, the K10 is long out, and while not bad, it is no where near as revolutionary as the K8. Of course the K8 was competing against the P4, which could be called a dog or an easy target.

If he can repeat what he did with the K8 (and if it was indeed due to his leadership), than AMD again has a chance. If not, they will end up like VIA and all the other Intel competitors, somewhere in a niche market.

> There are SO many ways that Samsung could have differentiated their products, but they chose to make it as similar to the iPad as they thought they could get away with. Other manufacturers havn't had any difficulty doing so. There are tablets in various colours, with textured non-slip backs, varying kinds of frontal designs. There were an almost limitless number of ways Samsung could have avoided this right from there start. But they chose not to. And now they're paying the price.

Exactly. Samsung went into what they knew was a legal grey area. The thing with grey areas is that sometimes you get away with it, but sometimes (especially if you are successful, and you are competing with Apple or Oracle) you do not. I have no sympathy with Samsung whatsoever

Great anecdote, but you could have just written "Motorola do not know how to do software". Never did, never will, well, unless Google changes things around there.

> I disagree with CDs and DVDs not being readable.

The CD-ROM is a stable format - it has been around for over 25 years already.

The die however may not last that long, depending on temperature.

Finding a CD-ROM drive in 25 years may be a challenge, flash memory way well have taken over by then.

And finally you also need to read the file formats. Common ones and standards like jpeg, mpeg2 etc should probably be fine.

> which is why they're pressuring Twitter to provide them again.

And if they have a legitimate interest, they should have no trouble getting a warrant. I think this is quite a simple case...

> Well, part of the problem is that people wear gloves -- and handle the same things they did before.

Yes, especially money. I have seen that many times, and I would assume it is always against company policy, and for good reason.

> Greenpeace sued to get the raw data, something I think they have a right to (since that study was used as the basis for approval).

And that's where the problem starts. Greenpeace may be well meaning, but they usually can't deal with data. And this inability is based on dogma, because data is seen as something "evil corporations" use to boost their profit (which they do, to be honest).

Working with data is not easy, and doing it from a dogmatic perspective (whether as a company or at Greenpeace) is never a good idea.

> "F.D.A. officials went to the inspector general at the Department of Health and Human Services to seek a criminal investigation into the possible leak, but they were turned down. The inspector general found that there was no evidence of a crime, noting that “matters of public safety” can legally be released to the news media."

Yes, that sounds to me like the bad guys are in control, and they are winning the battle. Stupid me thought that the FDA should be acting in the public interest...

> Windows XP is now 10 years old. It was released in 2002.

That is not the point. The point is that Windows XP was still being sold last year on netbook. To be honest, Windows 7 basic just does not make a lot of sense on netbooks, especially if you have only 1 GB of RAM. So there is no upgrade path, and in any case it would not be worth it.

People are not going to replace brand new hardware just because you tell them to.