Service interval in percents counter-intuitive. Why?

[san-ho-zay]

If using service intervals in percents, this is what the test for service required looks like in 0.5.1:

Code: Select all

v->reliability < GetEngine(v->engine_type)->reliability * (100 - v->service_interval) / 100

For example, if the max reliability of a vehicle is 90% and I set my service interval to 10%, the vehicle gets serviced when the reliability drops by 10% of the maximum reliability, i.e. 9% to 81%.

I guess this seems counter-intuitive to me. It seems more natural to have the percentage the other way round, i.e. 90% producing the same result as above. Was the idea to make it work the same way as the number of days service interval, i.e. a larger number means less servicing?

And, while I'm on the topic of servicing, can anyone tell me if the behaviour of automatic servicing at depot (no orders and no depot force) has changed from TTD? It used to be that trains would decide to go to a depot at the most inconvenient time in their route. Now it seems they make in intelligent choice. Where's this in the code - I can't see it in vehicle.c? Is it something to do with pathfinding? I was going to update the wiki page on servicing.

[PouncingAnt]

For example, if the max reliability of a vehicle is 90% and I set my service interval to 10%, the vehicle gets serviced when the reliability drops by 10% of the maximum reliability, i.e. 9% to 81%.

Gads, is *that* how it works?? I also found this counter-intuitive... in that I didn't know it worked this way around! whooops ^^;

[PandaMojo]

No wonder I felt the need to turn off breakdowns.