I disagree. If we don't have Net Neutrality, then the current advertising is deceptive and fraudulent. If we do have Net Neutrality and a real best effort to address network congestion rather than use network congestion as a payola scheme then there would be no need for the Federal Trade Commission to step in and put a stop to fraudulent advertising.

What high-level language does not depend on the C standard library and so would be suitable for implementing the C standard library?

It is possible, but for good performance it needs hardware support. We've implemented hardware-enforced bounds checking for C code using our processor. If you only care about accidental bugs and not about a malicious attacker, and don't use threads (or are happy to bound every pointer store with a transactional region), and don't mind that the semantics of C are subtly broken in the kinds of permitted pointer operations, then Intel's Memory Protection Extensions will do the same thing.

The OpenBSD philosophy says that the difference between a bug and a vulnerability is the intelligence of the attacker. There are lots of categories of bugs (null pointer dereferences, integer overflows) that were thought to be unexploitable, right up until someone exploited them. It's the same as with cryptosystems: the fact that you can't break your encryption algorithm doesn't mean that it's secure.

I remember when hard drives didn't fail. It was the time in between getting my first hard drive and it failing...

Actually, come to think of it, my first hard drive (40MB!) didn't fail, but my second one (60MB in a laptop) did.

Your laptop has to be on the same network as your backup machine, but even backing up my laptop over WiFi only takes a couple hours for an incremental backup. I don't have to leave it doing nothing, I just need to leave it on. If I haven't backed up for a while, I might leave it doing the backup overnight, but most of the time I run the backup while I'm working.

I bought 3 2TB disks just before the flood. About a month ago, they finally became cheaper than I paid. I'd been planning on swapping them out for 4TB disks after 2-3 years, but the 4TB ones are still 50% more than I paid for the 2TB disks. At this rate, 4TB flash will hit the £50 mark before 4TB hard disks...

The difference in cost between tapes and disks hasn't changed much, but the difference in cost of the tape drives to disk drives has changed hugely. You used to be able to get a tape and a drive for only a little bit more than the cost of the disk it would back up. It made sense to use tapes for backups then, because you could afford one tape for the same cost as a backup disk and add new tapes for very little money. Now, if you buy a disk at the sweet spot for price, the tape drive that can back it up to a single tape will cost you about an order of magnitude more than the disk drive. At that point, unless you want a lot more than 10 backups per disk, it isn't worth it.

That's not necessarily true. You can get the same amount of space in a smaller number of tracks around the edge of the disk, so the horizontal movement for the largest seek is going to be smaller. Seek times on mechanical disks are based on three factors. The first two are related: the time it takes to move the head between tracks (proportional to its distance) and the time it takes for the head to settle and be able to be lowered again (dependent on its speed). The third is the time it takes for the correct sector on the track to spin under the head. In the middle, you have fewer sectors per track, so you need to move the head more often (this is where the upper bound on seek times comes from).

It's probably the ultra-restricted version of Jazelle that is part of ARMv7 - it supports the branch to Jazelle mode instruction, but throws a trap when it is encountered so that it can be emulated by a JIT compiler.

Not a false dilemma, a false assumption. How do you demonstrate the proposition that a car without manual override controls (beyond just a big red stop button) is both individually and systemically safer than one without?

And what driver is going to wait and see if the car is capable of avoiding an accident if they are going to be the ones liable for that accident and told they must take control of the vehicle if they think the car is about to collide with something? Basically you are talking about taking a sophisticated collision avoidance system and short circuiting that by telling the driver they must take control of the vehicle if they think there is a problem. That could demonstrate that autonomous cars are less safe because people will be turning off the collision avoidance system at exactly the wrong times, but yet they will be reacting more slowly than people without autonomous cars because manual drivers are already actively driving.

What California is doing is starting from the assumption that not having a manual override is less safe, which I believe is a false assumption and actually undermines safety efforts. And it could also undermine efforts to roll out these cars.

I am all for the option of manual controls and would probably choose to have manual controls for a car that I owned, but I think that the more compelling case and safer option will be to remove the manual controls and I think the only way you prove that is by allowing the cars to demonstrate the capability.

Many of the most potentially beneficial things that could happen as a result of autonomous cars are those use cases where a driver isn't always at the wheel ready to take immediate control of the vehicle. Car sharing, taxi services, elimination of drunk driving, transportation for the disabled, highway driving at closer spacing which might make a human operator uncomfortable and prone to take control, congested city driving where vehicles could be routed and dispatched more efficiently or just told to "go park and come pick me up in twenty minutes" are all use cases where you don't want to require that someone is 'at the wheel' at all times.

With the real potential for saving lives and helping improve quality of life robot cars should be allowed to prove themselves with and without old school manual controls and all the legal requirements, increased costs and liability that retaining those controls imply.

Windows $[4+4]

First you have to allow the new technology. Requiring a licensed driver be at the wheel ready to take control of the vehicle at all times is not allowing the new technology it is hobbling it and potentially undermining the most compelling use cases that will save and improve lives. Simply require that the manufacturer demonstrate the ability of the car to drive like any taxi driver would be required to demonstrate an ability to drive to receive a license.

I think California is playing it wrong and unsafe. I agree there needs to be a big red button on cars which brings the vehicle to a safe stop much like there is on passenger trains, but this move by California seems more like something pushed for by entrenched vested interests and not driven by safety considerations. Lives will be saved when we allow cars to go pick up people that can't drive, don't have licenses or don't want to drive themselves. The implication of this move is that a human driver is going to be responsible for the operation of the vehicle at all times. Rather it should be the manufacturer of the vehicle which is liable for any defects of the autonomous system when it is driving autonomously. And it should be an option moving forward, even a safety feature, to allow cars without manual driving options except for the big red button.

Compared with the track record for human drivers which is proven to be completely unsafe?