So, I like trains, and I'm British, and I thought I'd clear up a few things in this thread:
- Thameslink is not related to the London Underground/the Tube. It was actually a project a couple of decades back (which is now in the middle of a major upgrade programme) to connect two major London commuter lines via a (previously disused) tunnel going north-south through the centre of London (the benefits of this are twofold — it saves capacity at terminal stations as the trains don't have to terminate in London, and it provides more capacity for people wanting to go through London). It's owned by Network Rail, the government body/not-for-profit company/entity with a weird status that owns most of the other railway lines in the country, and the trains themselves are run by private companies operating franchises funded, specified, and generally micro-managed by the government. At the time the article was written, this company was First Capital Connect, owned by First. Now Govia Thameslink Railways run the franchise, with these services branded as just Thameslink (they operate other services too). I would be very sceptical of the claim above that it handles the majority of London's commuter traffic, as there are plenty of much busier routes unconnected to the Thameslink network.
- So this is a valid WTF, but some people seem to have misunderstood slightly. I suspect (though I don't know for sure) that to get enough accuracy, the GPS used is assisted GPS, ie with a fixed base station as someone ruled out above for no apparent reason. In tunnels there are beacons that communicate directly with the train (that article implies via the GPS, so I'm inclined to believe it), and it is those that seem to be failing. They didn't just forget that GPS doesn't work in tunnels ;)
- In most places across the whole network, the current means of opening doors at stations with short platforms ("Selective Door Operation", or SDO) is for the driver or guard (if the train is not driver-only operated, or DOO) to specify how many doors need opening (you often have to do this in some arcane method like using the control panel for the door behind the doors you want opening in the train, or other such things. Obviously this doesn't work if you're the driver!). Unfortunately there are issues with this. It is not uncommon for drivers or guards to be momentarily distracted and forget how long the train they're driving is, or to forget that they're at a station that requires this and just open all the doors. Since it is usually (but not exclusively) smaller stations that require SDO, and sometimes it's only required on a few trains per day, it's not as hard to forget as you might initially think.
- Aside from this, there are issues with the number of doors that require opening being different on different trains (due to different carriage lengths), and for some trains to have really inflexible methods of selective door opening (I'm looking at you, Siemens!) requiring a different number of doors to be opened. So the general confusion caused by all this has made some view the current, manual method as unsatisfactory.
- Before you say "why not lengthen the platform or shorten the train", bear in mind that there are only two types of station that generally don't get platform extensions — those used so rarely that it can't be justified financially, and those that have some obstruction preventing lengthening (a bridge, a tunnel, a railway junction, a sharp curve (current health and safety laws prohibit new platforms being curved), etc.). In one case they had to close a railway just to lengthen a platform (Farringdon on Thameslink, the City Widened Lines that provided a route from Farringdon to Moorgate for surface trains had to be closed). And you can't shorten the train because people actually use trains in this country, and the whole reason the train was lengthened in the first place was that it wasn't long enough! We don't just arbitrarily say "Hmm, we want twelve coaches on these trains" — there are (or at least should be) calculations on the number of people who will actually be using them...
- For all these reasons, there has been recent experimentation with automatic SDO such as this. Unsurprisingly there are teething problems. But hopefully in the long term it will make the whole network safer, more reliable (through fewer accidents/incidents), and more flexible. Hopefully!
So what are the real WTFs here? I'd say the main one is implementing an unproven technology on a major route through the centre of London...