Atomic Systems

One of the most basic and fundamental of all train movements is a locomotive hauled train arriving at a terminus station, the locomotive uncoupling from the carriages, a light engine from a nearby siding coming and coupling to the carriages, that new train departing, and then the original engine running light back to the sidings or depot. In Rail Simulator, however, it is completely impossible for any AI train to do this, not because of just one failing, but because of a very large number indeed. I list them below:

1. An AI train cannot presently work with an uncouple order of any sort.

2. There is no possible way to stop the original engine from attempting to run back to the depot immediately after it arrives in the station, rather than waiting for the carriages to be pulled away, so it would crash into the carriages.

3. There is no way of telling a train only to start an instruction on the happening of an event, so there is no way to tell the engine from the depot coupling to the new train that it should not leave the depot until the first train has uncoupled.

4. The AI trains do not recognise stationary carriages, and cannot couple to them, but instead crash into them.

5. There is no way of changing the train class from light engine to express passenger or empty coaching stock, etc., when the new engine from the depot couples to it, nor for changing the class from express passenger, etc., to light engine when the original engine uncouples.

6. When the newly formed train leaves the platform, the signalling is such that a route could be set into the platform, even though it is occupied by a light engine coming out of the platform, resulting in a collision.

The result of all that is that it is quite evident that a very large amount of further programming work is necessary before Rail Simulator can simulate even the most fundamental of railway operations. A very large amount of programming work takes a very, very large amount of time, and we know that (1) RSDL has significantly fewer resources for continued development than Kuju had for initial development; and (2) with Kuju's greater resources, they were wholly unable to come even close to delivering on signalling, routing, pathing and timetabling despite the product being over a year late.

It is also apparent that Rail Simulator is unlikely to be a commercial success until the fundamental operational deficiencies are fixed, and that RSDL can only make enough money to cover its continued development from revenue generated by sales and add-on content from Rail Simulator. Unless, therefore, they are already a long way towards fixing all of the fundamental flaws and, inexplicably, do not want to tell anybody about it (which, all things considered, is overwhelmingly improbable) there is very real doubt indeed whether RSDL will survive for more than a very short period of time, leaving Rail Simulator flawed to the point of being nearly worthless.

Fortunately, for me, none of this is important. I just want an AI train which will run from A to B, look and act reasonably realistic - like stopping at red signals and station stops, and not disappear as soon as it reaches its destination unless its a portal.

I honestly think you are expecting too much from this software. It isnt, and i dont think ever was, intended to be a network simulator where you can simulate all real world train movements automatically. Its still essentially a train driving simulation which is all i want from it.

Lad491 wrote:Fortunately, for me, none of this is important. I just want an AI train which will run from A to B, look and act reasonably realistic - like stopping at red signals and station stops, and not disappear as soon as it reaches its destination unless its a portal.

That may be so for you and some others, but there are a very great many people indeed interested in proper operational functionality.

I honestly think you are expecting too much from this software. It isnt, and i dont think ever was, intended to be a network simulator where you can simulate all real world train movements automatically. Its still essentially a train driving simulation which is all i want from it.

It most certainly was intended to be a network simulator: in the pre-release publicity it was expressly advertised as such. The following is a direct quote from the Rail Simulator website as it appeared pre-release:

"Timetabled scenarios are more complex. A timetabled scenario is based around maintaining a whole network. With the option of jumping trains you can play for hours at a time going up and down the route trying to keep all the trains running smoothly and on time. If you get behind the knock on effect will be felt across the whole network!"

There can be no doubt whatsoever that Rail Simulator was always intended to be - and expressly promoted as being - a network simulator. That is what was supposed to have set it apart from the highly limited Train Simulator and Trainz, neither of which are realistic in their operational aspects.

It is also quite clear that there are fragments of functionality in Rail Simulator that only make any sense when used as a network simulator, timetabled scenarios themselves chief amongst them.

Finally, it most certainly is not asking too much: implementing proper railway operations is a matter of comparatively simple logic. Rail 3d, a programme written by a lone amateur in his spare time, is easily able to do all of the six things described above that Rail Simulator cannot do, although it has vastly inferior graphics and no sound.

I would be happy for a train to run from A-B ....trouble is they seem to run from A-?????? wherever they like!!!!!

As for the way you have to set up the destination by flying to it first.....very realistic in the sense that you may as well fly to your destination and not bother using the train....it kinda doubles up as a helicopter sim I suppose.
I was begining to like RS but if this is how you have to create scenario's with the results so unpredictable.

Abysmal

I agree that KRS could have simplified the routing system greatly if they had only emulated the prototype as I posted here - http://forums.uktrainsim.com/viewtopic. ... 20&t=56442
Although the system of routing points according to train code is not as intuitive as MSTS or KRS, IMHO its robust, simple yet flexible and can be used easily after a little practice.

With the option of jumping trains you can play for hours at a time going up and down the route trying to keep all the trains running

But the talk of "jumping trains" very much implies a driving simulator. Yes you should be able to input a 24 hour timetable if that is what you want, with trains starting at their correct times and running to their destinations. You can then jump off one train and into another, thus switching an AI train to a drivable train, and vice versa. But not expect the AI services to carry out those sorts of functions you outlined, possibly at remote locations far from where you are.

I still think you are reading into the description much more than was intended.

I read into it would be at least the equal of Zusi, which presumably the devs played at some point. Now granted you don't physically run round the train, but the programme will change locos for you or ends on a DMU etc.

That's what I had hoped for!

Lad491 wrote:But the talk of "jumping trains" very much implies a driving simulator. Yes you should be able to input a 24 hour timetable if that is what you want, with trains starting at their correct times and running to their destinations. You can then jump off one train and into another, thus switching an AI train to a drivable train, and vice versa. But not expect the AI services to carry out those sorts of functions you outlined, possibly at remote locations far from where you are.

I still think you are reading into the description much more than was intended.

That doesn't make any sense: if (1) there is a whole network of trains any of which one can choose to drive (which are otherwise by default driven by the AI); and (2) the services that one is able to drive have realistic operations, then it necessarily follows that AI trains must be able to perform those same realistic operations.

In other words, in a timetabled scenario, if one is to have the ability to jump to a locomotive hauled train arriving in a terminus, and then another light engine coupling to the carriages, etc., the AI trains must be able to do all that themselves since the user will get to pick and choose on the fly which trains are driven and which are AI controlled by train jumping. Note also the repeated use of the word "network".

Hi James,

I think you over-estimate how many people really care about such things. The vast majority of customers are quite probably Dads who buy a copy for their sons as it is cheaper than buying a train-set for Christmas. It is true that a few of the people on the forums will be keen on the fine detail of train movements, but how many of them even run scenarios.

The vast majority of users I hear from are more interested in building routes, building models etc. Activity (sorry Scenario) builders really are a minority in a minority hobby. My prediction is that after the first patched version (probably to coincide with the U.S. release), we will be very lucky if we ever hear from the developers again, so unless the community can fix RailSim like they did with MTS, then that is it.

Personally, I really like this simulator, I like the editor, and once we get 3D Canvas plug-ins available, I am hopeful of great things.

mikesimpson wrote:Hi James,

I think you over-estimate how many people really care about such things. The vast majority of customers are quite probably Dads who buy a copy for their sons as it is cheaper than buying a train-set for Christmas. It is true that a few of the people on the forums will be keen on the fine detail of train movements, but how many of them even run scenarios.

The vast majority of users I hear from are more interested in building routes, building models etc. Activity (sorry Scenario) builders really are a minority in a minority hobby.

The market that you describe is the market who already use MSTS and/or Trainz heavily, and thus cannot be interested in operational realism, since operational realism is completely impossible with both. There are also a large number of people who use SimSig and Rail 3d for whom operational realism is very important, and undoubtedly a large number of people who have not yet become interested in computer railway simulation precisely because there is no product that simultaneously has realistic graphics and sounds and realistic operations.

My prediction is that after the first patched version (probably to coincide with the U.S. release), we will be very lucky if we ever hear from the developers again, so unless the community can fix RailSim like they did with MTS, then that is it.

As things currently stand with the code, the community most certainly cannot fix it. If the issues are not addressed, Rail Simulator will be nearly worthless. It will also be a product that is vastly different (and many orders of magnitude inferior) to the product that was originally conceived of and promoted before release. Even as it stands now, Rail Simulator is described on the website in a way that is manifestly and demonstrably untrue (see the section signalling in "about Rail Simulator").

I can't see the point of building a route if you can't create the simplest of scenario's to run on it.
I have had three people new to rail simming who have been asking about RS for months and have been waiting to see what happens with the tools as they wanted to create their own scenarios etc. After seeing how you go about it and the results it produced they have decided not to buy it.
I think if the dads are thinking of buying it for their kids read the reviews on Amazon they will be putting up baseboards and laying track over Christmas!

So far I have found the activity editor to be easier to use than the MSTS one.
Dump the train where i want it, add a driver, set it as AI, set the passenger pickup points, set start time and i runs.

Couldn't give a monkey about timetables scenarios, right now all i want ois 4 trains running in a looop, and I managed it until I tried to change the start time

[quote][Dump the train where i want it, add a driver, set it as AI, set the passenger pickup points, set start time and i runs./quote]

There you go you have just created a scenario!!

mikesimpson wrote:Hi James,

I think you over-estimate how many people really care about such things. The vast majority of customers are quite probably Dads who buy a copy for their sons as it is cheaper than buying a train-set for Christmas. It is true that a few of the people on the forums will be keen on the fine detail of train movements, but how many of them even run scenarios.

The trouble is that the people who do really care about such things are the ones who are essential to take the game forward and ensure its longevity. You're absolutely correct though, the vast majority of purchasers won't care one iota for such detail such as this but most of these have probably forgotten about the game already and moved on to "Kill em' all 5" or whatever the current No 1 game is.

RS simply has to keep the a fair proportion of the hardcore on board and that's what concerns me. Far too many of the key players who got and kept MSTS 'up and running' with Routes, Activities and Models, seem to have almost given up on RS already.

I'm ploughing on with the RS Route I've been building for some time now, I'm just hoping that by the time I finish the track and scenery, that some of the talented guys on here will have sorted out the Signalling and Scenarios, so that I can actually use it in anger.

jamespetts wrote: That is what was supposed to have set it apart from the highly limited Train Simulator and Trainz, neither of which are realistic in their operational aspects.

Just to correct you there, the allegedly 'highly limited' Trainz can do everything in your initial post.

Mike.