JamesMoody
Trainz Team
Isn't the actual purpose of real rail roads to pick up something (where something could be lumps of coal, drops of liquid, pieces of metal, people, ... ) and take it to somewhere else?
Yup, and to do so as efficiently as possible (modulo a certain amount of "not invented here" that railways around the world seem to have).
Once you are past a specific scale, railways are doing more picking up and dropping off than one loco and one driver can cope with. And hence there's a lot of other trains rushing about.But I'd have to agree that doing only that can be a very unrewarding. Hence what has become my obsession with AI interactions which, at least to me, add an important element of reality to sessions and an extra layer of complexity.
The interaction between trains (not just in terms of waiting for one another to clear a specific point, but also in terms of interchanging cars) is part of the interest of the session.
The key thing here though, is via the session rules, the session must manage the interactions. It shouldn't be up to the player to /guess/ when it is safe to cross a specific pinch point, or to go to the next loop. In a real railway, there is generally a dispatcher or controller figure - and if you're tasking the player to drive a train in this kind of environment, with lots of other traffic, the session should play that role. This can be done through the session rules, setting junctions and displaying messages when appropriate.
A long time ago I had great fun with a suburban network - trying to path all those AI trains into the city terminus, at short headways. I think I had something like 18 trains per hour arriving and departing on the local lines, which gives a headway of a little over three minutes. This isn't just tight for the AI - it's tight for humans tooIs it even interesting to just use the 4-key view and watch the AI work their way around? Is it more interesting or have I been wasting my time?

All of those trains were pathed into the terminus onto different platforms, and had different places to go to at the 'country end'. They would all return to the city at the correct time later on to take up their position in the sequence again.
However, I think it was more fun for me making it than it would be to watch. I'm not sure the challenge of knitting all those paths together down one track would quite come across to a casual observer.
The key to it will be to integrate a human into the sequence, and demonstrate in the session via session rules what happens if you don't drive to path, with knock-on delays and so on. Turn up early, and there's stuff in front of you, and you have to wait. Turn up late and you get pushed from pillar to post as the signaller desperately tries to find somewhere to put you where you can't cause too much more damage.
Depends what you call a session. A text description in the session config file and a few wagons placed down in the session layer simply doesn't cut it. If you are building one properly, with the session rules doing such tasks as detection of rule violations, playing the role of dispatcher, and checking the player completes the task you have set, then no, it's going to take a lot longer than that.Finally, how long to make a session? Surely it more like an evening's inspiration, one evening to set it up + one more to test + a final evening packaging it all up?