Re: Y2K Fears to Stop CityRail (NSW) Trains

["Bill McNiven" <wmcniven@gunzel.ozemail.com.au>]

Eddie Oliver wrote, inter alia, in message
<37A63C1B.C1791A41@efs.mq.edu.au>...
>
>This assumption that all computer clocks know the exact time is just so
>removed from reality as to be laughable. The ones which people do know
>about, and have been checked for accuracy, are obviously the ones for
>which Y2K compliance will have been checked. So almost by definition,
>the ones which can create problems are the ones which people don't know
>about, and which will therefore probably not be accurate.
>

This thread has made interesting reading.  A few thoughts ...

First, most aus.railers probably don't go out on New Year's Eve - they're
all sitting at their keyboards flaming at each others speling and
punctuation.  So it's not an issue for aus.rail.

Second, I take Eddie's point that many hardware computer clocks are not set
correctly.  So if there is a problem, it doesn't necessarily come at
precisely midnight (i.e. AEDT, GMT + 11:00 in Sydney).  Even if the clock is
right, problems may come at 00:00 GMT.

Third, it's a reasonably safe bet that major systems such as power and
signalling (and Tangara speedos) will have been tested exhaustively.  (But I
wonder about the barrier indicators at Town Hall ...)

Some problems will come from little embedded process control computers in
devices such as fire alarms, burglar alarms, lifts, sewerage pumping
stations and the like, especially (as Eddie points out) if nobody knew a
computer was there.  Whether the date-handling is in software or hardware,
there are probably a few little computers in which the year after 99 is 00
(not a problem unless that implies dividing something by a time interval
like -3,155,673,659 seconds or its overflow equivalent).  There will be some
where the year after 99 is ':0' or '**' or similar.  Some will die and some
will malfunction, but not all at exactly midnight (or even on exactly
01/01/2000!).

The reason, I think, to close down CityRail at midnight on 31/12/1999 is
CROWD CONTROL. There will be big crowds at Circular Quay and Kings Cross
(NSW, Aust.) to celebrate the beginning of the last year of the millenium.
Back on 26/01/1988, Circular Quay station only managed by shutting the crowd
out as soon as there was a train load of people waiting on each platform.
More recently, Olympic Park (NSW) only functions by forcing a large crowd
(e.g. the 38,000 from the Manchester United match on 17/07/1999) to queue
outside until there is space on the platforms.  Even at times when most
customers are sober, an overcrowded city station offers the dangers of
people falling onto the track or stampeding on the stairs.

I've had to get home from some of Eddie's devious New Year's Eve mystery
tours (and my own), I've driven a cab on New Year's Eve in the distant past,
and I've travelled by train to the city for the fireworks as recently as
31/12/1997.

The demand for transport is DEAD around 23:30 to 00:15.  People who have
gone out somewhere get there well before 23:59 and they don't leave until at
least 00:15.  From a taxi perspective, the busiest period is 01:00 to 05:00.
On 31/12/1997, I travelled to Circular Quay around 23:00.  The train was
empty.  The street outside the station was full.  So were many of the
people.

At least for 31/12/1997, CityRail ran the normal weekend afternoon
frequencies with 8-car sets until trains leaving Central around 03:00.  This
is not "minimum headways", but is a significant train service.  I don't
think this service was needed 23:00 to 00:15, rather I think it was a case
of "keep the trains moving". The midnight fireworks finished at 00:10 and
significant crowds appeared at Circular Quay station from about 00:20.  The
peak at Central appeared to be around 01:15.

So, if the City Circle and the ESR shut down from 23:45 to 00:15, few people
will notice.

If all goes well, the service will be up and running when the rush comes at
00:20.

If something outside the railway goes wrong, the crowd will be outside the
railway too.  The most significant failure for the crowd at midnight would
be failure of cash registers in pubs, which would produce an ill-tempered
rush on rail transport. If some traffic lights, burglar alarms or fire
alarms fail, there will be a rush on emergency services so that "unhappy
crowd at railway station" is not their top priority.

If the cash registers and traffic lights fail at 00:00 AEST or 00:00 GMT or
at random intervals, it won't have the same effect on a crowd as a failure
at 00:00 AEDT.

Even if nothing fails, the hype about Y2K will make the 'customers'
restless.  Is it possible that people who usually meekly submit to
'trainsarerunninglateandoutoftimetableorder theindicatorboardhasbeenchanged
cityrailapologisesforanyinconvenience' will take this a proof of millennial
calamity and either stampede down/up stairs or set out to walk home through
the City Railway?

The discussion on power system restarts was interesting.  The last time I
was in the industry, Eraring and Liddell power stations had gas turbine
installations which would allow a black (i.e. no power coming in) start in
around 30 minutes.  (This is not a 'cold start' - it's akin to relighting a
steam locomotive when you've accidentally dropped the fire, but you still
have steam pressure.  The only catch is that you can't fire with a shovel.)
If there were a general power failure, it would probably mean an hour's
interruption to CityRail.

Incidentally, the p.m. service on 31/12/1997 was a one-off special timetable
and one expects that 31/12/1999 will be similar so "changing the rosters" is
not an issue.  Providing a 30-minute gap in trains through the city could
provide some nice crib breaks for crews at suburban terminals.

Regards
Bill

Disclaimer:  the writer has occasionally made a quid out of Y2K panics.  But
then, would you want a maintenance system that cancelled all time-based
inspections because they were years (precisely 100) overdue?