I think he was more likely referring to the two subsequent comments which pretty much slagged off a conference that might do some good to the debate.
I certainly wouldn't be a fan of your actual suggestion, either. The idea of predicating the purchase of something with a whole bunch of other completely unrelated actions doesn't sound sensible to me. The works you are describing would be simply punishing to anyone who didn't want to do them, and they would buy on the black market as they do now. I can understand the things you're prescribing as alternative kinds of punishment for minor crimes, but be honest about the fact that they ARE punishments. I don't think viewing the access to dope as a crime/punishment cycle is an improvement. It's actually what we have right now. You're talking about people getting pre-punished. Not a good idea IMHO.
How did this conversation end up sounding like a couple of stoned geeks?
I think I'd have to read the Android source to find out if the Location.GetSpeed() method uses the Doppler info.
Heh, yes, a naughty app for using on airlines when you're supposed to turn all that stuff off. I haven't tried any mid air tracking. There's apps for it already, so it's clearly possible.
But in practice not many chipsets / interface layers pass through the raw data to do that, so you’re stuck with whatever the manufacturer deigns to implement.
Yup, on Android all of the above mentioned interfaces are available, although a particular phone might not present them (although why it would bother to have the hardware at all in that case, I don't know). There are a few more as well. Magnetic compass readings, humidity and pressure sensors, the list goes on. All adds up to the chance of a better reading, with the right software. I just trolled through some of the developer doco, and most people seem to do it off absolute location smoothed over previous readings, which is what I recall the developers of the software I use saying they did for the most part. It's hard to argue with "you've gone this distance in this time, hence this is your speed".
Software specifically designed for boy racers used the accelerometer more, because boy racers want to know their true horsepower, their 0-100km/h and their standing quarter mile (fuck that anywhere but a race track in my last muscle car which got to 100km/h in 5.5s - I shudder to think what it would have got to by the end of a quarter mile), and they really need to know the g forces for some of that. Also cornering g-forces really only come from the accelerometer. One app I have makes a little movie of it all overlaid on the view out the window. Strikes me as the absolutely perfect thing to use if you really want to get busted to the maximum limit the police can manage, as they watch your collection of geocoded clips shortly before crushing your car.
Intriguing, didn't know that was available. Software can use any algorithm, taking into account multiple sources - cell tower, satellites, and accelerometer, and possibly the map knowledge of your probable altitude (in case you've only got 3 satellites locked in), all together and smoothed. Doppler shift is probably quite accurate for working out velocity, although as with the accelerometer, I'd think that waving around isn't going to help matters, if the polling interval is slow.
GPS doesn’t always give an accurate speed reading – ever had a look at your GPS track if using a fitness app like MapMyRun? It can be out by a few metres here and there, which must have an impact on your speed calculations, so GPS is not as infallible as many people believe
It's certainly not very accurate for an instantaneous reading. But for a reading over time, since it is based on absolute location, and is smoothed out by the accelerometer (which gets very confused by being waved around as you run), it's far more accurate than something counting the number of times your wheels turn. In a vehicle going 27 meters per second, with a polling interval of a few seconds at a time, and the "meter here meter there" averaging out to the exact location, you've got an excellent doublecheck on your speedometer (and also, for that matter, the odometer). You'll definitely be able to detect it being out by 10%. The app I'm talking about gives you continuous feedback on how many satellites it is using, so you know when it's getting a really good reading.
Part of being a good driver is being aware of distance and speed – if you are having to use an app to keep under the speed limit, then you are probably not sufficiently aware of your speed, and you should be paying more attention to the task at hand.
My point was that our speed can rove up or down 5km/h at 100km/h within perfectly normal driving (in Auckland, which has hills). The only way NOT to do that is to micromanage your speed, braking and accelerating at every little deviation, or sit well below the limit.
In terms of the actual need to use the app, I only every use it on long motorway trips. It's very easily to lose concentration on the speedo, since you're focusing on the road. I can't see it being a negative to have an app that helps. It would be nice to say that I never ever, ever forget to glance at my speedo every 5 seconds, but that's not realistic.
Droid, soz. Could be on iPhone, it's called "Ulysse Speedometer". There's lots of these types of apps.
Speedometers are normally calibrated/designed to overread – this is apparently required in AU and the EU.
It's one of the cool things about GPS. You can test your speedo quite easily. I have an app that hassles me if I'm going over 100km/h on my phone. It's interesting to try to drive to that - it gets really, really irritating if you set it to actually go off on 100km/h, and then want to drive at 100km/h. It goes off half the time. The only way to avoid being nagged by it is to go 95 on average. Or set it to 105. Quite instructive about what does actually happen with your speed on average - we don't micromanage it, on the whole, nor should we.
It’s the millions of New Zealanders who gained their licenses in previous decades and have never since been tested on their abilities that I’m more concerned about.
I think refresher testing is not a bad idea, but accident statistics suggest that there's not going to be much bang for your buck on the older demographics, who have by far the lowest accident rates already. It's not really the knowing of the law that's the issue in bad driving. It's the obeying it, including the hard-to-enforce ones like following distances, and taking poor conditions into account. And also the general improvement in overall road sense that steadily builds from endless repetition, close shaves, minor accidents.
Is there another dangerous machine that you would you put people in charge of without ever reviewing their ability to use it?
Unfortunately, yes. There are currently no laws requiring licensing of any kind to drive extremely powerful small boats, nor any manditory seaworthiness testing for the boats, and there's very little by way of DIC controls. And people die frequently as a result. Which is not an argument not to improve road safety, but is a strong one for improvement of sea safety regulations.
If the tests are relevant to driving, is it possible that people who suck at that sober shouldn’t be driving even when sober?
That's always been the rub. You couldn't make the test too hard, if it was the decider, people would be failing it left right and center. But it's just a filter, aa way of narrowing down to who to do a chemical test on. For that purpose, it sounds efficacious, although still a hell of a lot slower than getting me to say my name into an alcohol sensor. The manpower requirement would be huge.