Slashdot Mirror

because the odds against someone noticing something as subtle as a defective ABS module during a post-crash investigation are astronomical.

Compare the effort that goes into investigating a rail accident (example -- a derailment with no injuries, a 41-page report concluding with learning points and recommendations) to the not-much that happens after a road accident.

I'd like to see more effort spent investigating road accidents -- perhaps choosing a small sample to thoroughly investigate.

Imagine the costs involved in treating every auto accident like a plane crash!!

But imagine the benefits!

The UK has the safest railway in the world (of any reasonably sized country). Look at the rail accident investigation reports -- pick one -- to see why. The reports usually finish with "recommendations already implemented". (I sometimes read these reports, typically the ones about structures or vehicle faults -- the causes are often very technical, and it's interesting to read about how they work out what caused the accident/near-miss).

(This is probably expensive. We also have one of the most expensive railway systems in the world, but it's not clear what the cause is. I don't know why we don't have the same focus on road accidents -- globally, British roads are in the top 5 for safety, but they're still poor compared to railways.)

The only close call I had was a train crossing with a malfunctioning gate that I didn't realize was malfunctioning (wasn't down, wasn't even blinking) until I was doing 45 about two carlengths from the tracks and so was the train

That's really poor on the part of the railway. Did they investigate how it happened?

Something similar happened in England in 2011. The independent investigation is 57 pages long (PDF): http://www.raib.gov.uk/cms_resources/111215_R202011_Lydney.pdf
Evidence from a "black box" on the train, the signalling system, and a data logger for the level crossing itself was used to work out what had happened, how it happened, and why. Most importantly, they recommended what should be done to prevent it happening again.

The result of this kind of attention to safety and risk is the UK has the safest railway network in the world (ignoring tiny countries). An accident (or "near miss") isn't shrugged off, but investigated to try and reduce the risk of a recurrence.

(An ironic result of this: in 2010-11, a single railway worker died while at work. He was an infrastructure maintenance worker, and was killed when the van he was in crashed on the A9 (main road) in Scotland. For 2012-13, there were two fatalities. One was killed on a motorway on the way to work.)

Automatic cars should have the same approach to safety. We could almost eliminate road deaths this way.

Trains should also be very safe, which may well make the Canadian accident the more interesting story.

I read some of the RAIB (British railway accident investigation) reports, as they're published. They're usually things done during maintenance (workers putting a trolley on the wrong track that's still in use; letting a piece of rail-wheeled equipment runaway) and rarely about normal freight or passenger trains -- the routine stuff is very safe.

http://raib.gov.uk/publications/current_investigations_register.cfm

All actual work is done in metric.

Is it? A couple weeks back I read an RAIB incident report (in a discussion of why trains take so long to stop, a mention was made of a train that overran its intended stopping point by two and a half miles, and I was curious about the details), and I noted that they used the following units:

Track positions were measured in miles/chains, long distances were measured in decimal miles, and short distances were measured in meters.
Speeds were measured in miles per hour.
Flow rates were measured in kilograms per minute.
Accelerations were measured in %g.
Short times were measured in seconds, while long times were measured in fractional hours.
Pressures were measured in bar.

Actual work is done in metric, you say?

The "last chance to disconnect" point would have to be tens of kilometres from the end of the parallel track.

You absolutely have to allow for exceptional scenarios, such as when "a passenger train running from London Charing Cross to Hastings failed to stop at Stonegate station in East Sussex. The train ran for a further 2.45 miles (3.94 km) with the emergency brake applied, passing the level crossing at Crowhurst Bridge before coming to a stop 3.22 miles (5.18 km) after first applying the brakes.". That train was only going at 100km/h and took over 5km to stop. It was due to high winds causing exceptional autumn leaf fall, making the rails exceptionally greasy, and a maintenance oversight meaning there wasn't enough sand to give extra adhesion. (investigation report)

I notice the guy with the idea is a designer, not an engineer. (I'm not an engineer either, FWIW.)

I sometimes read the reports from railway accidents in the UK -- there's often interesting science (physics, metallurgy, etc) involved in establishing what happened, why, and how to prevent it happening again.

The reports always give many causes. For example, this one (where a bridge over a river partially collapsed) had these causes:
- The bridge was undermined by the water in the river (immediate cause)
- An obstruction in the river channelled the flow towards the east abutment
- The railway company was not aware of the obstruction
- Staff inspecting the track were not aware that they must check for obstructions in rivers crossed by railway bridges
- There was no way for members of the public to report the obstruction
- The bridge (from 1858) had shallow foundations
- ...lots more

Without investigation we won't know what happened in this case, but if the man died by impacting the steering wheel (etc) then not wearing a seatbelt is likely to be a causal factor. If he was already dead then it obviously isn't.