🚀 go-pugleaf

ok@goanna.cs.rmit.OZ.AU (Richard A. O'Keefe) sez in <3155@goanna.cs.rmit.OZ.AU>
>In article <2371@skye.ed.ac.uk>, ken@aiai.ed.ac.uk (Ken Johnson) writes:
[Ken writes about a Prolog database detailing the structure of the
London Underground]
[Richard points out faults]
>It seems to me that a better structure for this data base would be
[map info]
>	line(FromStation, ToStation, LineName)
>	-- you can get from FromStation to ToStation in one step
>	-- by travelling along the line called LineName

Probs: one part of the Central Line looks like this:

                                South Woodford
                       ----- *-----*--...
        Leytonstone   / Snaresbrook
...--*-----*---------<
   Leyton             \ Wanstead
                       ----*-----*--...
                              Redbridge

In the database this looks like:

line(leyton,leytonstone,central).
line(leytonstone,snaresbrook,central).
line(leytonstone,wanstead,central).
line(snaresbrook,south_woodford,central).
line(wanstead,redbridge,central).

However if you want to get from Wanstead to Snaresbrook you have to
change at Leytonstone.

Or look at Epping, where there's a capacitor-like symbol.   Or Monument
and Bank, which are connected by escalator.

The Underground has quite a few special cases waiting to foul up such a
program.

Such a program will need a concept of "Epping trains" and "Hainault via
Newbury Park trains" and "Peak Hours only" and "Westbound" in order to
be really useful. 

(A few thoughts from a Londoner who lives in South Woodford and so has
to think about these things.)
-- 
\/ o\ Paul Crowley aipdc@uk.ac.ed.castle                       :-| Have a day
/\__/ "The whole is the sum of its parts, plus one or more bugs" aimsm@castle