Re: European train deaths

[Matthias Andree <matthias.andree@gmx.de>]

Colin_Paul_Gloster@ACM.org (Colin Paul Gloster) writes:

> Something which occured to me only yesterday (unless I cleanly
> forgot before) is that Ada and formal methods are used for
> European train systems and so these may be involved in some
> of the spate of fatal crashes over the last circa two years
> in the U.K.;

Nope, AFAI have information about UK train accidents, at least one (near
London in 1996 or when that was) was caused by a driver's fault. 

In Germany, the Deutsche Bundesbahn has - among other systems "InduSi"
and "SiFa", systems that prevent a train from entering a "block" (track
segment between two signals) which is red. This system was not in place
or working properly in at least one of the accidents in the UK.

If the driver fails to stop the train, the train will stop anyways
(several signals transmitted inductively), this cannot be overriden by
the driver. Should the driver fall asleep, die, leave his seat, the
train will trigger an emergency stop procedure c. 30 s after the
incident. I don't know if speed limits can be enforced in a similar
manner, I think they can, see below.

As to some of the major German train accidents, on the big ICE disaster
in Eschede, there was a mechanical defect. In a recent other major train
accident in Western Germany (was it Emmerich? Not sure) the driver went
much too fast through a detour (which went around track constructions or
maintenance), with 120 km/h with 40 km/h allowed. In Wuppertal, where
the Schwebebahn dropped off its tracks and fell into the Wupper river,
workers had forgotten to remove a track clamp which is used to prevent a
train from proceeding through a track construction.

In Belgium (which has three official languages), there was a
communication problem since the driver did not properly understand the
warning issued to him.

I believe there's nothing Ada could do about these incidents. I cannot
tell about the Eschede disaster since I believe the last words aren't
yet spoken on that case.

-- 
Matthias Andree