The International Simutrans Forum

Development => Extension Requests => Topic started by: Václav on August 05, 2011, 04:19:19 PM

Title: Change of pre-signals behaviour back to v. 102
Post by: Václav on August 05, 2011, 04:19:19 PM
We, who are inside Czech language forum, are a quite confused from current behaviour of pre-signals and we would like to see behaviour of pre-signals as it was in called version - it means train waited until next two fields are empty, not that trains wait until all fields are empty.

Else solution would be in preparing of next type of signals - with such behaviour as pre-signals had in version 102.
Title: Re: Change of pre-signals behaviour back to v. 102
Post by: prissi on August 05, 2011, 07:22:42 PM
Can be easily arranged. But there are more commetns need on this ...
Title: Re: Change of pre-signals behaviour back to v. 102
Post by: Václav on August 05, 2011, 08:55:28 PM
Thanks. Welcome to Czech language board. There you can get needed informations.
Title: Re: Change of pre-signals behaviour back to v. 102
Post by: prissi on August 05, 2011, 11:05:32 PM
I need other player comment. In english. Czech is not readable for me, sorry. And I need some arguments too.
Title: Re: Change of pre-signals behaviour back to v. 102
Post by: Víťas on August 06, 2011, 05:55:55 AM
Here is my experience.
I didn´t download 110 until today,because I´ve read some arguments that the system of signals in 110 is not good. Today, I downloaded 110, I loaded my map with a lot of pre-signals and a lot of trains and...all the system got stuck  :-[ . So if the behaviour of pre-signals won´t change, I play 102.
Víťas
Title: Re: Change of pre-signals behaviour back to v. 102
Post by: Václav on August 06, 2011, 07:54:39 AM
Quote from: prissi on August 05, 2011, 11:05:32 PM
I need other player comment. In english. Czech is not readable for me, sorry.

And I need some arguments too.
1. of course  :) - but I waited that you would use any translator  :D

2. what type of arguments do you wish?
Title: Re: Change of pre-signals behaviour back to v. 102
Post by: jamespetts on August 06, 2011, 10:12:27 AM
Prissi wants an argument? Try here:

http://www.youtube.com/watch?v=kQFKtI6gn9Y (http://www.youtube.com/watch?v=kQFKtI6gn9Y)
Title: Re: Change of pre-signals behaviour back to v. 102
Post by: Ters on August 06, 2011, 02:09:34 PM
I don't get this. I have been using simutrans straight from svn for some time now, though I am a few days behind now. As far as I can see, my presignals reserve two blocks ahead, while normal signals reserve one. Isn't this what they have always done, or has this changed in the last few days? Or maybe I missed some change/bug some time back that has been undone?
Title: Re: Change of pre-signals behaviour back to v. 102
Post by: Václav on August 06, 2011, 03:07:09 PM
Ters, and what version do you play?

Quote from: jamespetts on August 06, 2011, 10:12:27 AM
Prissi wants an argument? Try here:

http://www.youtube.com/watch?v=kQFKtI6gn9Y (http://www.youtube.com/watch?v=kQFKtI6gn9Y)
James, could you tell me what is go in that video? My English is not so good I could understand that video - and what is more important, Monty Python is not cup of tea that I would drink.

...
and as this discussion goes, I am getting to be thinking more and more that signals need a little greater diversity
Title: Re: Change of pre-signals behaviour back to v. 102
Post by: jamespetts on August 06, 2011, 03:22:52 PM
That's the famous Monty Python argument sketch.
Title: Re: Change of pre-signals behaviour back to v. 102
Post by: Ters on August 06, 2011, 05:27:04 PM
I don't play a version, I play revision 4721. That might change during the day, though.
Title: Re: Change of pre-signals behaviour back to v. 102
Post by: TurfIt on August 06, 2011, 06:46:38 PM
I don't get this either. AFAICT presignals are working as intended and the only changes since 102 have been bug fixes. What specific behaviour do you believe has been changed?

One of the bugs fixed was signals after a presignal were always treated as a regular signal. If a chain of presignals was built, in 102 only every second would be functional, in 110 the train will wait for the whole chain to clear.

I can't read Czech either; The online translators do a horrible job so I'm probably lost in the translations. All I could get from the Czech board were references to yellow lights. Presignals are not 3 aspect signals (Red/Yellow/Green). All signals in Simutrans are 2 aspect only.


Quote from: Víťas on August 06, 2011, 05:55:55 AM
Today, I downloaded 110, I loaded my map with a lot of pre-signals and a lot of trains and...all the system got stuck  :-[
Can you make this map available?
I have recently encountered trains running a red presignal upon a save/reload but haven't been able to isolate the cause yet...
Title: Re: Change of pre-signals behaviour back to v. 102
Post by: prissi on August 06, 2011, 07:12:37 PM
Actually presignals are three aspect signals, when the first block is free but the second is not.

One could treat a signal following a presignal as normal signal for driving on. However, if this signals is a choose signal or long block signal, its function needs to be check; otherwise the train would be stuck where according to the presignal it should not stuck.

Changing this behaviour to not check the other signal, if this is a presignal would not require much code change, but could lead to situations where trains stuck on places they are not suppose to stuck.
Title: Re: Change of pre-signals behaviour back to v. 102
Post by: TurfIt on August 06, 2011, 07:31:47 PM
Quote from: prissi on August 06, 2011, 07:12:37 PM
Actually presignals are three aspect signals, when the first block is free but the second is not.
???  presignals have the train waiting at the presignal until both following blocks are clear. A three aspect signal from real life signalling would display a yellow aspect and the train would proceed up to the next (red) signal.


Quote from: prissi on August 06, 2011, 07:12:37 PM
Changing this behaviour to not check the other signal, if this is a presignal would not require much code change, but could lead to situations where trains stuck on places they are not suppose to stuck.
Possible but I think keeping presignals as they current operate is best. The only use for them is to ensure a train doesn't block a junction. I would hate to introduce cases where they don't work for that intended purpose. Especially after my pending work trying to keep road vehicles out of intersections!

Anybody have a track arrangement where such chained presignals are necessary/useful?
Title: Re: Change of pre-signals behaviour back to v. 102
Post by: Václav on August 06, 2011, 08:15:56 PM
Quote from: TurfIt on August 06, 2011, 06:46:38 PM
in 110 the train will wait for the whole chain to clear.

Can you make this map available?
1. it is what is go - on corridors only pre-signals should be built (or not?) - and when such corridors are very long, trains don't go - because they have to wait until chain is clear; I combine it with red signals but ... those ones should not be in chain

Trains should wait on pre-signals only for next two fields ahead to be clear, not all. So, it is reason for what I wrote in in my previous post - that it is time for a little greater diversity of signals.

2.  ??? I don't understand. What map?
Title: Re: Change of pre-signals behaviour back to v. 102
Post by: Lubak91 on August 06, 2011, 09:43:48 PM
Presignal in version 102: http://www.youtube.com/watch?v=OX20UkT3ZrM
Presignal in version 110: http://www.youtube.com/watch?v=COq2S7M4YXM
Title: Re: Change of pre-signals behaviour back to v. 102
Post by: prissi on August 06, 2011, 10:01:59 PM
This is not the right use for presignals. It just works because of a bug.

Show us a savegame please, where the signals are used correctly but new behaviour causes problem.
Title: Re: Change of pre-signals behaviour back to v. 102
Post by: Lubak91 on August 06, 2011, 10:11:27 PM
Prissi:
Presignal in version 102: http://www.youtube.com/watch?v=OX20UkT3ZrM - The track can hold a lot of trains => good :)
Presignal in version 110: http://www.youtube.com/watch?v=COq2S7M4YXM - The track can hold a very few trains => bug :-[
Title: Re: Change of pre-signals behaviour back to v. 102
Post by: Ters on August 06, 2011, 10:28:07 PM
Why so many pre-signals in a row? Wouldn't normal signals work just as well? Pre-signals in Simutrans are not really like real pre-signals. To me they are just a necessary evil to be used only when normal signals may cause deadlocks. Maybe the name should be changed to avoid confusion.
Title: Re: Change of pre-signals behaviour back to v. 102
Post by: TurfIt on August 07, 2011, 12:18:15 AM
Quote from: prissi on August 06, 2011, 10:01:59 PM
This is not the right use for presignals. It just works because of a bug.

Show us a savegame please, where the signals are used correctly but new behaviour causes problem.
Ditto.

The tracks shown in the videos should be using standard signals.
Title: Re: Change of pre-signals behaviour back to v. 102
Post by: Václav on August 07, 2011, 06:04:53 AM
Quote from: TurfIt on August 07, 2011, 12:18:15 AM
The tracks shown in the videos should be using standard signals.
...
I have ever used them (exlcuding pre-signal before end of corridor)
...
May it be it is not correctly as it was planned in Simutrans - but right as it should be (based on real world) - because Lubak91 is one who understand signals most of us.
Title: Re: Change of pre-signals behaviour back to v. 102
Post by: Víťas on August 07, 2011, 04:06:24 PM
Here you are.I´ve created a new map,because the old one was quite big with lots of czech addons.
Try it in 102, where it works correctly and then in 110.
http://vitasovyvs.wz.cz/Save.rar (http://vitasovyvs.wz.cz/Save.rar)
Title: Re: Change of pre-signals behaviour back to v. 102
Post by: prissi on August 07, 2011, 06:30:46 PM
Apart from using new graphics, so I can not tell what of them are actually presignal by looking at them ... You aparently used presignals for normal signals. That is not allowed and worked only for a bug.

And before version 90.0 you could run as many trains as you like without using any signal, due to a bug.

Presignals garantee that a train entering the next block can leave thing block (for trains shorter tahn signal spacing) - maybe you understand something different by this. But this is the definition and purpose of presignals in simutrans.
Title: Re: Change of pre-signals behaviour back to v. 102
Post by: Václav on August 07, 2011, 09:10:42 PM
... I think that here is a little misunderstanding caused by:
1. difference between czech signals and signals from country that was used as pattern (regardless of I think it should be the same - but I am not sure)
2. effort to make signals as close as is possible to real world.

So ... why don't prepare another type of signals what would be functional (and used) in way that is discussed (it means train will wait for two empty fields, not for all) - and red signals could be used elsewhere - where their using would be more suitable.
Title: Re: Change of pre-signals behaviour back to v. 102
Post by: prissi on August 07, 2011, 09:28:36 PM
This signal exists: It is the normal signal for such purposes. If you do not want to use it, then do not complain that it is broken.
Title: Re: Change of pre-signals behaviour back to v. 102
Post by: Ters on August 08, 2011, 04:45:32 AM
Quote from: VaclavMacurek on August 07, 2011, 09:10:42 PM
I think that here is a little misunderstanding caused by:
1. difference between czech signals and signals from country that was used as pattern (regardless of I think it should be the same - but I am not sure)
...

I don't think the behavior of pre-signals in simutrans is patterned after any country. They work as they do because sometimes that behavior is necessary to overcome shortcomings (or whatever the proper word is) in simutrans' simple routing and signaling. In the real world, timetables and human signallers make sure trains don't enter blocks they can't leave. Am I right?
Title: Re: Change of pre-signals behaviour back to v. 102
Post by: Václav on August 08, 2011, 10:33:45 AM
... this topic was not my idea ... and I end it.

Of course that I have some ideas what could be made - but about them whenever else.