My visit to RSC today

[thecanadianrail]

And now there's a real reason to comply with speed limits: if you go too fast for the condition of the track, you'll derail even on straight, plain line.

Paul

[bpetit]

Cab sway: was it totally, utterly 100% real-life realistic? No. Was it the silly side-to-side seasicknss of RS in the early days? No. Are they still working on it? Yes. I'd give it 8/10 at this stage. What is really cool is that unlike BVE, it isn't random and hard-coded into the core. Cab sway is driven by two factors: the track rules, and the characteristics of the train's suspension - both configurable by the content creator. So if you take a train with stiff suspension over a really well-maintained piece of track (eg HST out of Paddington), then cab sway is hardly noticeable. If you take a Geep over a yard that hasn't seen maintenance in years, then it bucks and sways. And now there's a real reason to comply with speed limits: if you go too fast for the condition of the track, you'll derail even on straight, plain line.

Paul

Cool! Thanx Paul.

[OldProf]

Thanks for the information, Paul. Since I like writing scenarios with interactive AI traffic, can you tell me whether you saw any new AI action (hint: AI coupling, rather than just uncoupling)?

Tom Pallen

[mikesimpson]

Thanks for that Paul, but do you think 23rd September really is set in stone? It is only about 9 weeks away and yet you say much of the development is still needing to be done (you mention multi-core CPU optimisation and adding the new graphics engine to the World Editor as 2 items not yet implemented) - These are pretty major upgrades and I know from experience how dangerous it is to announce release dates for software. (Remember how many times Microsoft were going to release Vista? )

Mike

[Trainguy76]

Mike, I think someone posted at UKTS (who visited RS.com) said they had left lots of time to get everything ready for release. Plus, it's not entirely an update, but set up as a "Release" of Train Simulator 2012, I think they would try their hardest not to botch it up.

[jp4712]

I don't think there's any such thing as 100% certainty in this kind of thing, but Paul Jackson was very confident. I think it's important to draw the distinction, as Paul did in Chatham, between the six-man operation when RSC was formed and when everyone had to do everything, to the company now with 25 people and where the work is much more process-driven and in a stable environment. The optimisation they talked about was about making the code more efficient before it goes out, loose ends really. I think (but do not know) that what we saw was already running multi-core. A date was mentioned for code-freeze which clearly I'm not at liberty to mention, nor the budgeted coding time that Paul said is in the plan to take place beforehand, but the date we were given was somewhat in advance of release date so I think they have probably assumed some time for sweeping up and beta testing before September 23rd.

@Tom: we did not see AI interaction but as the RSC guys have already 'spillied the beans' in public I can say that there is at least one big AI step forward: which is that an AI service will be able to couple up to, and take away, a consist so long as a) it is in a marker such as a siding or platform; and b) that consist has been there since the start of the scenario. So it won't take away a train that you, or another AI service, have just brought in: but it is a very big step forward. Incidentally this is a feature that some people shout about, but given that AI dropping off has been possible for some time, how many scenarios actually show it? Not many.

Paul described the AI pick-up behaviour as 'one step forward, not the whole journey' so I'm confident that more is to come. For example see viewtopic.php?p=24164 and check out what Adam says about dispatcher versioning. Read his words well: what he says is much, much more than smoke and mirrors.

Paul

[imnew]

Hi Paul and thanks, very nice read. One question for you, or one word is more correct, "stuttering". You know that RW can stutter even though you are running high FPS. How did that come out on your visit? Did you notice smoother gameplay, runbys and so on?

[PapaXpress]

A question about repainting. Will repainters benefit from any of the new enhancements in TS 2012? Should we put our projects on hold until then?

[Trainguy76]

A question about repainting. Will repainters benefit from any of the new enhancements in TS 2012? Should we put our projects on hold until then?

The only thing I know that will be different is the lighting system, which *may* make the colors look a little different.

[jp4712]

One question for you, or one word is more correct, "stuttering". You know that RW can stutter even though you are running high FPS. How did that come out on your visit? Did you notice smoother gameplay, runbys and so on?

You can't imply anything from what we saw. It was a lash-up dev version (hybrid of RW2 editor and RW3 in-game), before the final code optimisation. Yes it was pretty smooth, but we weren't given a print-out of the PC specs it was running on. However, I don't think I'm going out on a limb by saying that multi-core processors are intended to be used more effectively in RW3 and people who have them should be checking to see what improvement they get. I did see stuttering as you entered a tile in view '2' a couple of times, but as I say you can't draw any conclusions at this stage.

Paul

[imnew]

Im guessing/wishing that when all is ready for release, stuttering will be a thing of the past.

thanks for the reply Paul, appreciate it.

[Hawk]

I wouldn't get my hopes up that the stuttering will disappear completely.

[jp4712]

No, nor me. Like I said, you can't read anything into what we saw, given that it was an alpha-level dev version. I'm going to go out on a limb here, and say that in RW3 if you have an i7 quad-core processor with liquid-cooled GPU and solid-state HDD, you will probably get less rubber-banding and stuttering than a standard laptop.

Paul

[Hawk]

That's a long limb Paul.

[jp4712]

Hey, I'm on a roll! While I'm there, I also predict trouble in the Middle East at some time over the next 100 years.

I'm surprised that no one has asked me more about projecting lights? What we saw was, in the World Editor, something resembling some sort of Child Object. This disappears in-game. It appeared to me that the intensity of the light seemed to be configurable by the content creator (or at least the person who creates and places the child object) so you can provide very powerful headlights or less powerful ones. I surmise - but do not know - that this opens the way to retro-fitting projecting light sources to DLC. So, to explain: it seemed to me that the 'light source' that we would all point at and would say 'look, there's the light' was not the projecting light source, but that the child object that fitted 'over' the headlight and which was not visible in-game was actually the source of projected light.

I do not know if light color is configurable but the screenshots we've all seen seem to imply that they are. I also believe - but cannot confirm - that lights are available as projecting cones, or as a point projecting light in all directions.

Paul