Open Rails does not follow player path file

[lateagain]

Thanks for that Link Charles, that's useful. Hopefully the ET problem can be resolved soon.

Hi Chris, I realised the signature was supporter, just wanted to know if any of the opinions expressed reflected an "official response".

Sadly those of us who've been around since MSTS arrived have seen a number of other Sims come and go and every time this happens there is a proportion of the Train Sim Community that seems to find it necessary to "nail their colours to a post" and declare that anyone who doesn't do so is some sort of luddite. Any comment or criticism of any newer Sim seems to be taken as a personal affront? It's a sports team supporters outlook as far as I'm concerned.

ANY development in Train Sims has to be welcomed by true Train Sim fans and the idea that anyone who's stuck with MSTS would not welcome a new Sim with backward compatibility is ridiculous, AS IS the idea that this backward compatibility is not a HUGE advantage.

It's nearly 50 years since I gave up a secure job and went to Art School and learnt design. The first things you learn are that there are many people with MUCH greater talent than yourself and more importantly you learn the value of constructive criticism. Criticism is sometimes hard to take but if it's valid and you address the points raised in that criticism your design is all the stronger for it?

An interesting parallel to this was shown a couple of days ago on TV in the UK (doubtless on BBC America soon?) looking at Cars. James May's "Cars of the People" http://www.bbc.co.uk/iplayer/episode...es-2-episode-1 looked at how the US and UK car industry collapsed because they didn't listen to criticism or look at what their competitors were trying to achieve.

I for one don't want OR to end up like Detroit or British Leyland?! LOL.

[NorthernWarrior]

Interesting thread to follow, not just because of the subject matter - but the insight into the OR development psyche. Always a bit sceptical of this type of development path (looked what happened to OpenBVE) as there is always the possibility that like a house buying chain, one (or more) people pull out or lose interest and the whole thing is in jeopardy. So if I sometimes come across as less than engaged with OR, the "man-cave" metaphor kind of sums up why.

On the matter in hand, well MSTS had zero dynamic routing capacity, trains followed the player path to the letter and would not deviate. The only commercial sim where I have seen this (and randomisation of routing) is Zusi2 where, at the timetable creation stage, the author can specify what routes are available to a train at diverging switches. Kuju failed to code that logic into Rail Simulator and DTG have failed to make any advance on that situation in the latest TS2016 code. So any improvement OR can bring to the table in that respect is, I guess, welcome though the full power can only be achieved once OR has its own Activity Editor and alternate paths can be specified.

[conductorchris]

Geoff and Vern have some valid concerns that have crossed my mind too. I am grateful for - and impressed with - what has been achieved in open rails. Yet there seem to be some structural weaknesses and unmet needs (not just editors, but ongoing marketing to bring in new blood).

But what to do about it? I'm not a coder. Nor do I have a lot of spare time. Some of these weaknesses just seem to come with open source projects or hobby efforts in general. It's not like the other simulators are what I'd call strong either.

Until we have a strategy, I'm going to spend my emotional energy on doing what I can to support those who contribute so much. Because their contributions are appreciated.

Christopher

[R. Steele]

Mr Steele, Are you part of the Open rails Development team and if so what part do you play in it?

No I am not, but you've got that figured out already.
Mr. Brown, please call me Gerry.

Thanks for that Link Charles, that's useful. Hopefully the ET problem can be resolved soon.

Hi Chris, I realised the signature was supporter, just wanted to know if any of the opinions expressed reflected an "official response".

Sadly those of us who've been around since MSTS arrived have seen a number of other Sims come and go and every time this happens there is a proportion of the Train Sim Community that seems to find it necessary to "nail their colours to a post" and declare that anyone who doesn't do so is some sort of luddite. Any comment or criticism of any newer Sim seems to be taken as a personal affront? It's a sports team supporters outlook as far as I'm concerned.

ANY development in Train Sims has to be welcomed by true Train Sim fans and the idea that anyone who's stuck with MSTS would not welcome a new Sim with backward compatibility is ridiculous, AS IS the idea that this backward compatibility is not a HUGE advantage.

It's nearly 50 years since I gave up a secure job and went to Art School and learnt design. The first things you learn are that there are many people with MUCH greater talent than yourself and more importantly you learn the value of constructive criticism. Criticism is sometimes hard to take but if it's valid and you address the points raised in that criticism your design is all the stronger for it?

An interesting parallel to this was shown a couple of days ago on TV in the UK (doubtless on BBC America soon?) looking at Cars. James May's "Cars of the People" http://www.bbc.co.uk/iplayer/episode...es-2-episode-1 looked at how the US and UK car industry collapsed because they didn't listen to criticism or look at what their competitors were trying to achieve.

I for one don't want OR to end up like Detroit or British Leyland?! LOL.

You've misinterpreted my posting. I was reminding people to "work the problem" , which in this thread was "Open Rails does not follow players path" not side issues about where the OR developers hang out, the mechanism of bug reporting, etc. etc.

Geoff (may I ), I would definitely agree with almost all of your post #31, the exception being:

"...proportion of the Train Sim Community that seems to find it necessary to "nail their colours to a post" and declare that anyone who doesn't do so is some sort of luddite. Any comment or criticism of any newer Sim seems to be taken as a personal affront? It's a sports team supporters outlook as far as I'm concerned."

That attitude may exist (I do not doubt your experience), however, it is not reflective of my attitude. Generally I support OR because it is an Open Source volunteer project, an admirable pursuit for the community. Also I genuinely want the project to succeed, and one way for that to happen is for users (like us) to come together and rationally discuss the pros and cons of Open Rails - on its own merits not using a 15 year old incomplete program as the yardstick to measure it.

Although, to be honest, I am really ignorant when it comes to most anything having to do with railroading. I have always been fascinated by trains and now is when I have the time to pursue it - in a digital form. Open Rails, to me, appears the best way - for me -
to do that. (although I will always recommend RUN8 to people, it does provide another unique experience)

Okay, now I have gone dashing off into the weeds, my apologies to geepster. Sorry

Work the problem. AND above all else have fun, this is only a hobby, just a small part of life.

[csantucci]

In the Feather River activity mentioned earlier in this thread, the player is sent onto the diverging path with nothing coming in the opposite direction. This happens with or without location linked passing selected. Far as the player path is concerned, this should be followed, but it seems that OR is treating an AI running in the same direction the same as if it was running in the opposite direction. Instead of holding the player train at a red, it uses the alternative path into the siding which has a red at the far end because the exit is still blocked by the preceding AI.

Has the alternative path you mention been defined as a passing path in the player path?

[jpantera]

Geoff and Vern have some valid concerns that have crossed my mind too. I am grateful for - and impressed with - what has been achieved in open rails. Yet there seem to be some structural weaknesses and unmet needs (not just editors, but ongoing marketing to bring in new blood).

But what to do about it? I'm not a coder. Nor do I have a lot of spare time. Some of these weaknesses just seem to come with open source projects or hobby efforts in general. It's not like the other simulators are what I'd call strong either.

Until we have a strategy, I'm going to spend my emotional energy on doing what I can to support those who contribute so much. Because their contributions are appreciated.

Christopher

Am currently working on a personal project using the MSTS NWC route. I have built a timetable with around a 100 services and am adding more in regularly. OR is giving me a lot of the features that other Sims dont. I am right behind it but find when asking questions on forums I rarely get a reply. I have then instead of giving up figured out workarounds or fixes myself.

Have been pointed in the direction of ET forum but came up against a closed door. A massive amount has been achieved with OR I hope it continues and will support it while it exists. A lot of folks have questions as they want to get the most out of it.

As for my project I will continue to persevere and hope that the developers continue with what so far is the best Trains Sim experience I've found.

[baldwin]

Even if it is defined as a passing path, there is no logic in using a passing path with all traffic at that point in time travelling in the same direction.

[csantucci]

OK, with this info it's easier to track down the problem. But affirming that "Open Rails does not follow player path" is not exact and misleading in this case. The player train is following the passing path that has been defined within the player path.

[csantucci]

I tried to reproduce the situation with this attached packaged demo activity for the USA2 (Marias Pass) route with legacy trainsets.
The player train has its path defined to the siding, while its passing path leads on the main. The player train has a WP in front of the signal at the end of the siding.
The AI train behind it has its path defined on the main.

Running the activity (also in autopilot mode) the player train correctly takes the siding and stops in front of the signal. The AI train overtakes it on the main. Everything as programmed.

So in your application there seems to be a further "disturbing" parameter.

Testpassing.zip

[geepster775]

In my case with the Orin Line act, location linked passing path option, checked or not, had no effect.

When I re-built the player path removing the passing paths defined for that location, the player train finally responded in Open Rails by making the correct crossover move at Jerome (success), however, because it was not anticipating a meet per the lack of passing defined in the path file, I stopped at Osage facing a red restricting signal rather than a hard red absolute stop (failure). I got a restricting indication because the AI trains wait point was inserted before its approach signal for Osage. OR would have allowed me to incorrectly tab past the Osage signal and nudge up to a standoff with both trains on opposite sides of the approach signal on the single track. Clearly, removing passing points from the path is not a fix. They need to be in there for a reason. What needs to be improved is the application's path decision-making abilities when these passing points are present in the path.



Here are the 2 trains approaching each other, including their "sonar" (as I call it) that explores the situation ahead of the trains. The AI is at a dead stop per the wait point, but its red line already has gone through the control point at Osage and is occupying Main 2.




A close-up of the AI "sonar" already occupying Osage interlocking and a portion of Main 2 while the wait point clock runs down.




Since the AI claimed Main 2 between Osage and Jerome first, I am unsure as to why the player train was even allowed to deviate from the prescribed path and run the same Main 2.




Not until the AI's wait point expires and the AI begins moving again, does its sonar "flip" over to the diverging route and allow the AI to diverge - When it shouldn't be diverging per the original plan anyway.




I will also state that in the Feather River act, the AI train immediately ahead of the player train also has a wait point defined in its path, immediately in advance of Kramm siding. The first AI and second AI do not share the same path in that act, with the wait point being the only difference early on.

Open Rails seems to be delivering the most basic, rudimentary, standoff-free meet executions (what I call unintelligent meets, successful yes, but no prototypical realism in how it was carried out), while actively fighting against user-defined tweaks done to add the missing realism to the meet's execution. Some activities being designed today go beyond the application's predictive abilities. And in both activity failures, Open Rails is deciding to deviate from the defined player path when there is no cause to.

In my view, the player's service is not seeing ahead enough to detect that the AI was first to claim Osage interlocking and first to occupy Main 2 in a westward direction. Since the AI grabbed Main 2 first, and was being held by a wait point to give the illusion it was arriving last, the eastbound player service should not have had any option handed to it, except to stick to the original defined path and jump over to Main 1 at Jerome.

[OTTODAD]

I tried to reproduce the situation with this attached packaged demo activity for the USA2 (Marias Pass) route with legacy trainsets. The player train has its path defined to the siding, while its passing path leads on the main. The player train has a WP in front of the signal at the end of the siding. The AI train behind it has its path defined on the main.

Running the activity (also in autopilot mode) the player train correctly takes the siding and stops in front of the signal. The AI train overtakes it on the main. Everything as programmed.

Yes, it performed as you have designed it to do, Carlo !

Sorry about the terrain on tracks glitch, I did not bother to fix in this default route, using DB-tracks !

[geepster775]

Yes, it performed as you have designed it to do, Carlo !

Unfortunately, this example is not at all relevant to the problem(s) discussed.

In the Orin Line example, we are attempting to make the player train step aside for the oncoming AI, and Open Rails is not delivering because, for some unknown reason, possibly because of the extended length of the "siding" involved, it is deviating the player train from its path when it should not.

In the Feather River example, all 3 trains should stay on the main and run on each other's yellows without anyone using any sidings for overtakes, and Open Rails is not delivering because it keeps attempting to force the player train, which is bringing up the rear, off of the path and to take the siding when its unnecessary. Again, its attempting to deviate when it shouldn't.

Proving it works within the confines of this simple overtake example is not at all related to these 2 broken examples which involve more complex trackwork or lean harder on the application's predictive and thinking abilities. And strangely both "broken" acts run correctly in MSTS.

[csantucci]

OK, I made some confusion about the two activities (Orin and Feather River).

[geepster775]

I appreciate all the help.

Tonite I'll try the act again by removing the wait points in the AI and adjusting the timing of the AI so it doesn't need to stall, and see if the routing logic for the player responds differently. I often use wait points for AI traffic like this as a way to make the player train take a delay and leave a siding somewhere at a known time. Doing so reinstates predictable timing in the activity mid-way through it.

[OTTODAD]

A long time ago I worked on creating signalled activities, trying to simulate the BLOCK SIGNALLING systems in use on real railways / railroads.

Had to give up and became an apostel of Sniper Jim Ward's FREESTYLE activities. refraining from using AIs in them which are like to get into the path of its player or other AIs.

Lets face it, to me AIs are just candy to liven up the action and using the same idea in Railworks it does heaps better, allowing multiple players to be run, one at a time and multiple AIs to run on tracks not likely to be used by players, in one of my Freestyle scenarios up to 24 of them !

Those who prefer prototypical action in Train Simulators have to be satisfied with what is on offer, unless they can make limited improvements themselves !