Rail transport ("Tunnel vision", THES, December 13) has traditionally been viewed as the safest form of transport since its designers subscribe to the concept of "failure to safety", that is, when there is a problem, all potentially affected trains are stopped. This has led to a public perception that rail must be safe, at almost any cost. Airlines, with an equally impressive safety record, are allowed a set number of failures (usually viewed as acts of destiny). They are not required to perform exactly to schedules: the public does not want to land in fog or snow if this would present a risk.

The Channel tunnel fire has shown, again, that many serious accidents are caused after a system has failed (safely) due to a lack of reliability, in this case the position of detectors on the loading ramps. Tunnel infrastructure prevented deaths or serious injuries. Very often, though, back-up systems and written procedures can lead to catastrophe because they are inherently less safe than the system which "failed to safety".

John Adams argues that the Channel tunnel is, essentially, a waste of private and public money because its construction time was longer than predicted and its cost soared. However, he does not analyse the reasons for these problems. Many of the design changes and additions to the original concept were forced on the promoters by safety authorities and public bodies which had no direct involvement in the design process. The result was an inconsistent structure whose only claim to risk reduction was the strict adherence to "failure to safety".

Once the fire had occurred, the whole link was declared safe, that is all movements were stopped, even though two tunnels rather than a single (and cheaper) two-track tunnel had been built. A single-country supervisory authority would have insisted that one bore be reopened the day after the incident. The intergovernmental safety commission prevaricated, causing delays and negative publicity.

³ÉÈËVRÊÓƵ

The lessons from the fire are expensive and must be learnt:

* major infrastructure investment must be funded by the body benefiting from it: the community

³ÉÈËVRÊÓƵ

* system design should not be left to specialists, each of whom is only interested in their own area

* integrity of the system should be maintained when a component fails, that is, the system should degrade gracefully and not "fail to safety" regardless of the situation

* the integration of private-sector projects into the public environment requires new and better regulatory approaches.

Our society insists on mega projects and total solutions and it does not take a terrorist attack to cripple the system. Snow does the job quite nicely on the M25 and at Spaghetti Junction. When will the IRA go for these two targets?

³ÉÈËVRÊÓƵ

FELIX SCHMID, Director, MSc in rail systems engineering, University of Sheffield