OMG, KRS team have finally posted!

[johndibben]

1. Not working as we want it to work can be the equivalent of broken when "how we want it to work" is in a way that can realistically simulate railway operations, and "how it was designed to work" is in a way that cannot realistically simulate railway operations. If that is what is meant, then it can properly be described as "broken" even if it is working as designed. Not all flaws in software are bugs.

Not working as we would would wish isn't a flaw but the lack of something we would wish for.

It all depends upon the definition of the term 'we'.

[RSderek]

Posts read James, I'll attempt to clear the waters on a clear head.

regards

Derek

Think it may be time for another glenmorangie

[ordan77]

Think it may be time for another glenmorangie

Derek, if you can improve the signalling and I'll send you a bottle

[RSderek]

,

there was some text here but it has now gone.

[phat2003uk]

Derek, I recommend you edit your post before anyone notices .

[jamespetts]

Was considering signalling/pathing/timetabling and wondered why they shouldn't be the trains in preset positions or set up in portals with a set of instructions and no points set.

It is tricky and I've found following an activity in MSTS requires driving to positions with little margin gor error. The same applies with timing. Slightly out and activity failed. This is surely playing a game with perameters more harsh than reality.

Then there stops for signals and a wait for other trains which may or may not arrive. The waiting time is wasted. If it's required then a fast forward would remove some of the boredom and uncertainty.

Given that we often manually operate some points and pass others on red with permission, we deviate from being simply a driver and reality. We could set them all.

It's been said a simple signalling system can be worked out on a 1960's computer but as a fan of sigsims, I've found reality to be far more complex.

John, I wrote that because computers built in the 1960s actually did run real life railway signalboxes in the 1960s

Not excusing any bugs in KRS but wondering if a signalling system which fits all eventualities is a realstic goal?

It is far easier than realistic graphics, sounds or physics: as I have written elsewhere, the solo amateur who created Rail 3d made a simple (granted, not fully featured: no overlaps, for instance, but close enough), but very powerful signalling system in his spare time.

There are wider issues of gameplay.

As I understand it we can still 'chase the yellows' and conduct any move but pre-programming all these moves with AI isn't possible.

It's not about pre-programming every move with AI: it's about designing a system that is simple yet powerful that will mean that those moves don't have to be programmed with AI, but can be plotted using very simple logic. See here for details.

Remember, there are a great many people who are interested in doing far more than driving trains: the operational side is just as interesting as driving.

If we accept this is KRS and not MSTS then the issues appear less important although bugs would be better if fixed.

Rail Simulator was supposed to be just that: a railway simulator, not a simple train simulator. Simulating a railway requires simulating signalling just as much as simulating train physics.

[stewart]

Derek, I recommend you edit your post before no-one notices .

Too late, we've all seen it. Nah na na nah nah!

[RSderek]

If only every thing could be fixed with a paint brush, i'd be ya man.

right, bed time for me.

[jamespetts]

Not working as we would would wish isn't a flaw but the lack of something we would wish for.

It all depends upon the definition of the term 'we'.

If "something we would wish for" is realistic simulation of railway operations, and the thing that lacks it is a product that purports realistically to simulate railways, then "broken" is a fair comment, even if not everybody is bothered by the fact that the bits that are broken are, indeed, broken.

Broken things, however, can be fixed, and, since RSDL has promised to support the product in the long-term, there is hope yet.

Posts read James, I'll attempt to clear the waters on a clear head.

regards

Derek

Think it may be time for another glenmorangie

The reply is very much appreciated I certainly understand that posting detailed replies to complicated things at eleven something in the evening is not conducive to making a great deal of sense

I hope that you and the rest of the team appreciate that much of our frustration at the things that haven't been got right (yet) is in large part due to the fact that much of the game is very well designed and shows very great potential: I love the graphics and the lovingly recreated routes and the delightfully modelled assets, and the sounds and physics (which, although not perfect, are better by far than any other railway simulator of which I know), and very much hope that I will be able not just to drive trains up and down, but participate in a realistic set of simulated railway operations with a full timetable and realistic signalling one day using the generally rather good game engine that you have all been designing for the past two years or so. I daresay that I am not alone in that view.

I also appreciate that commercial pressures mean that you have not had as much time as you would like to produce the game as you would like it, and that you are (in the circumstances, very sensibly: rather a troubled product now which is fixed in time than no product at all) trying to compensate by building on the product post-release, with the input of the community. Although some of us wonder whether, even given the limited time, some aspects of signalling could have been done better, ultimately, as long as it is all made to work properly in the end, one way or another (where "the end" is not many years from now), I, at least, will be happy.

Incidentally, one aspect of the operational side that you all have got right (unlike, I think, all other railway simulations, unless Trainz has it and I didn't realise) is portals: although they don't currently work for player driven trains, when fully implemented, they will go a long way towards maximising the realism of operations over a slice of a larger network.

Enjoy your glenmorangie, and I shall look forward to any new information

[MartinvK]

...

Incidentally, one aspect of the operational side that you all have got right (unlike, I think, all other railway simulations, unless Trainz has it and I didn't realise) is portals: although they don't currently work for player driven trains, when fully implemented, they will go a long way towards maximising the realism of operations over a slice of a larger network.

Enjoy your glenmorangie, and I shall look forward to any new information

You mean those things where a train can enter one and reappear at a later time in either the same or a different one? Trainz has been there, done that long ago. They can generate AI traffic to add trains to a route to give it that lived-in , used look, so the view out the window is not of an empty world. They'll load or unload passengers or freight. Someone made a rerail-portal to put derailed trains back on the rails so you can continue your session. There is even a iportal version that can send and receive across the internet.

[AndiS]

It's been said a simple signalling system can be worked out on a 1960's computer but as a fan of sigsims, I've found reality to be far more complex.

John, I wrote that because computers built in the 1960s actually did run real life railway signalboxes in the 1960s

I need to say something in defence of Kuju here, for a change.

It is true that they have computer programs, and before relay boxes and electro-mechanical stuff that can be modelled by a few bits, do the train control, and it worked. There are a few things, though, which we forget.

A) For big stations, these things were pretty complex and designing and "coding" them required persistence nearly no gamer will have.

B) There were a lot of rules protecting these systems against undesired states and these rules were obeyed.

C) There were lots of humans in the system, in particular when shunting, doing a lot of things that electronic could not do then, most of all find out whether the path of travel is obstructed or not.

D) Most of all, this is a game. It needs to be sold to people who use it for their enjoyment in their spare time. These people want to rip apart sections of track, glue them together in some other way and start the train, 5 m from the "construction site". They cannot be bothered of even considering the impact of their changes on block devices, interlocking, track circuits. Most of them do not even know that such things exist.

This is what makes the job so hard in KRS. There are many things which would be needed to make the system run in a smooth and simple way, which are simply not existent. Instead, the signals try to figure out what the rest of the system would do if it were there.

[sp762]

*Group hugs*

Take a tip from me and act dumb.

It's worked for me over 17,000 times

Is there a point at which the act becomes reality, he wondered idly?

[KlausM]

Well, the fact that RS is buggy (e.g. the errors in the dynamic switches) and unhandy (mostly editor, but also such basic things like the limited "free roaming" view) in many aspects is annoying, but can be solved somehow. The problem I see is more fundamental and that's why I am quite reluctant at the moment: Up to now, the RS shows far less flexibility than announced, and it needs to be proved that Kuju/RSDL has learned its lessons from other years old simulators. Starting from such completely stupid things like placing buffer stops at each end of the track, obviously bound to the style of the track (can that be turned off somehow?), over the the whole catenary thing (which isn't worth saying anything more) to the signalling issues. While the graphics seems to be quite good as far as I can see it on my old machine, the railway technological framework is — sorry to say that — on a level of a toy train. My fear is that the release of the developer tools will only confirm that.

Klaus

[bigvern]

Not wishing to labour the point but the signalling system in Zusi is an example of how it should work in any other train sim.

It works equally well whether you wish to drive one train or "hop" between different services in the timetable.

It will "regulate" lower priority trains for higher priority, or sort out the effects of late running. It will also occasionally put in random diversion or replatforming.

The core of any train sim needs to be an autonomous signalling and control AI which will deal with any scenario or activity overlaid. Otherwise we have not advanced any further than the rigid fixed paths of MSTS and the frequent lock-ups they caused.

It's a relief to see RSDL acknowledging the concerns but rather than tugging my forelock in hero worship at a bit of banter, as someone who paid £35 for a game now sitting unused after just over a week due to the large number of flaws and bugs some real answers on the many issues (and when we might see a patch) would be more appreciated.

[mearle73]

I just hope they take on board what has been said,It could be a great sim,I dont want a pink 0-6-0 with gt stripes,all I want is the major faults fixed,or to be told it cant be done.