Page 1 of 1
| Joining and Dividing 17/03/2010 at 19:40 #868 | |
clive
2737 posts |
I want to draw timetable writers to an issue. Consider a situation where two trains arrive in opposite directions for a scheduled join. One of the trains (call it E) will have nothing more in its timetable. The other (call it C) will continue on with more activities and perhaps more locations. So we have, say: E: ... Shepreth Branch Junction; Cambridge J:C. C: ... Coldham Lane Junction; Cambridge J:E, DR:N; Shepreth Branch Junction .... The situation I want to refer to applies when C has a divide or detach after the join. When trains arrive in opposite directions to join, technical issues with the core code mean that one of them will reverse direction. If this is E, this won't matter. However, if it is C, this means that the subsequent divide will be at the unexpected end. In the above example, the meaning of the timetable is that the new train N should appear at the Coldham Lane Junction end. However, if C has reversed for these technical reasons, N will appear at the Shepreth Branch Junction end instead. (Note that C is now facing the wrong way to depart for Shepreth Branch Junction, but it will automatically reverse again to solve this.) In many circumstances, which train reverses is consistent. Timetable writers may have got used to this situation and put a DF:N instead of a DR:N to resolve it. In other circumstances, which train reverses may differ from run to run of the simulation. This bug has now been fixed in the core code. In all simulations released from now on [2], all divides and detaches will take place in terms of the arrival direction of the train, irrespective of whether it does a join first or not [1]. This means that timetable writers need not worry about this situation in future. HOWEVER, if you have found about and coped with this wrong reversal by experiment in the past, you will have to edit timetables. To reiterate, this situation only occurs when a train has a join followed by a divide or detach at the same location, and the train it joins to arrives from the opposite direction. [1] However, please note that at the first location of a new timetable, the "arrival" direction of a train is the direction it will depart in. So front of train N will depend on where it is going next. [2] But excluding any at the Peterborough meet. Log in to reply |
| Joining and Dividing 17/03/2010 at 19:40 #7336 | |
|
clive
2737 posts |
I want to draw timetable writers to an issue. Consider a situation where two trains arrive in opposite directions for a scheduled join. One of the trains (call it E) will have nothing more in its timetable. The other (call it C) will continue on with more activities and perhaps more locations. So we have, say: E: ... Shepreth Branch Junction; Cambridge J:C. C: ... Coldham Lane Junction; Cambridge J:E, DR:N; Shepreth Branch Junction .... The situation I want to refer to applies when C has a divide or detach after the join. When trains arrive in opposite directions to join, technical issues with the core code mean that one of them will reverse direction. If this is E, this won't matter. However, if it is C, this means that the subsequent divide will be at the unexpected end. In the above example, the meaning of the timetable is that the new train N should appear at the Coldham Lane Junction end. However, if C has reversed for these technical reasons, N will appear at the Shepreth Branch Junction end instead. (Note that C is now facing the wrong way to depart for Shepreth Branch Junction, but it will automatically reverse again to solve this.) In many circumstances, which train reverses is consistent. Timetable writers may have got used to this situation and put a DF:N instead of a DR:N to resolve it. In other circumstances, which train reverses may differ from run to run of the simulation. This bug has now been fixed in the core code. In all simulations released from now on [2], all divides and detaches will take place in terms of the arrival direction of the train, irrespective of whether it does a join first or not [1]. This means that timetable writers need not worry about this situation in future. HOWEVER, if you have found about and coped with this wrong reversal by experiment in the past, you will have to edit timetables. To reiterate, this situation only occurs when a train has a join followed by a divide or detach at the same location, and the train it joins to arrives from the opposite direction. [1] However, please note that at the first location of a new timetable, the "arrival" direction of a train is the direction it will depart in. So front of train N will depend on where it is going next. [2] But excluding any at the Peterborough meet. Log in to reply |