The fact that Railtrack did nothing to address this problem, despite repeated representations from the people on the ground, must constitute neglect in their duty of care for providing a safe railway.
The other issue that must be addressed is the automatic warning system (AWS). Despite sounding a hooter in the drivers cab (if it was working - it was not at Southall) when signals other than green are approached, AWS does not differentiate between double amber, single amber and red (the sequence the Thames driver should have seen). This has led to drivers forgetting which amber signal was passed and overshooting the danger signal.
Maybe the driver was expecting signal 109 to change from red as the train approached and didn't allow time to stop. This happened at the Colwich crash when the Euston to Manchester overshot the danger signal and derailed the Liverpool to Euston train.
Jason Barton
London
David Street's letter on Great Western Railway's use of AWS (October 8) gives the wrong impression of the system. AWS is in fact already installed throughout the British railway network, indeed it has been on most routes for over 30 years.
Basically, when a driver approaches either a yellow (advance warning) or red (danger) signal a hooter sounds in the cab. If the driver does not press a button to cancel this warning within five seconds the brakes are automatically applied. AWS is simple and effective but not foolproof - once a driver has cancelled the device there is nothing to ensure he then slows or stops the train. Indeed, most incidents where a signal has been passed at danger will have occurred despite AWS being in operation.
The train protection system recommended after the Clapham accident automatically reduces the speed of any train approaching a red signal should the driver fail to do so, ultimately bringing the train to a halt if needed This makes it impossible (in theory) for the driver to pass a signal at danger.
Jeremy Hinks
Newtown, Cheshire
I must correct your suggestion (The agony of those who wait, October 8) that the Thames train involed in the Paddington crash could not have forced the points open. These were "facing" points governing diverging tracks. A train can only force "trailing" points where tracks converge.
What has now emerged is that the points were actually set for the Thames train to cross the path of the Great Western despite the latter having been given a green signal.
This would have been impossible under the old Westinghouse mechanical interlocking system, for the clearing of the green signal for the Great Western train would have automatically locked the points for the Thames train in the ahead position. Thus the Thames train, despite having overshot signal SN109 at red, would not have collided with the Great Western, but would have continued straight forward.
Conversely, with the points set previously across the path of the Great Western, it would have been impossible to give this express a green signal. A Railtrack spokesman is alleged to have said: "The old system of directly linking points and signals has been phased out with computerisation".
Surely this interlocking system should be restored and inserted in the computer's programme as quickly as possible.
To permit the present system to continue to operate is criminally dangerous.
Brian M Poag
Beckenham, Kent
The automatic train protection system was deemed too expensive at £1bn. We are told that government actuaries calculated the cost per life saved of installing it at £14m (roughly 70 lives), and that was too high.
The cost overruns of the Jubilee line extension would have more than covered this expense. The government is finding that money, of course, in its frantic attempts to make the Millennium Dome accessible on time.
Working on the insurance principle that what has occurred will recur, presumably the number of deaths which would be averted by the ATP system will have gone up dramatically since Tuesday. Perhaps that will bring the cost per life saved down to a level that the government and Raitrack would regard as good value.
Jonathan Wilson
Oxford
I am currently working on one of the major housing developments in the south east of England. One of the requirements of the project is that the buildings are secure by design. The idea is to design out the problems as best as one can, and then utilise the technology to act as a final catch-all.
It would appear from the latest coverage of the Paddington disaster that a hi-tech approach is seen as the palliative to the problem of human error. Bi-directional running of trains is inherently less safe than running on double tracks. Head-on collisions on bi-directional track have happened before, and will happen again even if reliance is placed on technology.
Surely it would be better to design a safer railway, and maintain it properly than to rely solely on the questionable magic of a technological solution. The abiding rule for my work is KISS - Keep It Simple, Stupid. There has to be some merit in trying to apply this to the design of a safer railway.
Andrew Searle
London
andrew.searle@trenchfarrow.co.uk
While on my way to work this morning my train stopped at signals five times.
Before yesterday (when it stopped four times), it had, as far as I remember, never once stopped. What conclusions am I to draw from this?
Jon Robinson
Cardiff
