Friday Facts #68 - The trains

[cpy]

I was hoping for some train love like electric trains/tankers and stuff, but bugfix is cool too.

I'm hoping to see ships one day, sea biters, cruisers where we go! Offshore drilling rig (oil/metal?) hell yeah, space? Hell yeah. Factorio can only get better.

[fregate84]

What are pre-signals doing? some big signal network or so?

the pre-signals is used when you have 2 or more paths (to station for exemple). The pre-signals is red if all signals are red.
It is used to block train before the train have to choose to go to the way 1 or 2.

like that :

without pre-signals


with pre-signals

[ssilk]

Won't a path signal have significant amount of common code with a presignal anyway?

Pre signals are simple: the need to check, if some former signals have red. (Very simplified)

Path signals are quite different and much more complex. They look, which path the train will go (they look into the trains planned path) and "reserve" the rail segments, where another train can possibly collide. That enables to have several trains in one block, if the trains cannot collide.

[MF-]

Path signals are quite different and much more complex. They look, which path the train will go (they look into the trains planned path) and "reserve" the rail segments, where another train can possibly collide. That enables to have several trains in one block, if the trains cannot collide.

Thanks for the explanation.

So a path signal is just a "signpost" that tells the train to reserve it's path via GSM-R or something.
The logic of that is no longer situated in the signal device.
I don't like that.

PS: Also it means the train needs to know where it's going AND don't change it's mind in the protected (super)block.
With the current planner, I don't think that's the case.

[MF-]

What are pre-signals doing? some big signal network or so?

the pre-signals is used when you have 2 or more paths (to station for exemple). The pre-signals is red if all signals are red.
It is used to block train before the train have to choose to go to the way 1 or 2.

like that :

without pre-signals


with pre-signals

To describe in words: "Never stop in the marked (super)block -- between presignal-start and presignal-end

[MF-]

Won't a path signal have significant amount of common code with a presignal anyway?

Pre signals are simple: the need to check, if some former signals have red. (Very simplified)

Is the following example a downside of presignals,
or a reservation-in-advance needs to be implemented for presignals as well?


You might still need to do the "Reservation-in-advance" thing,
so another train from an auxiliary track won't block your free station.

Example of that:
There are two stations, just like in the picture above. I'll call them s1 and s2.
They're guarded by presignals.
However, the block of s1 can be reached from an auxiliary track as well.
That's outside of the presignal (super)block.

When a train see s2 full and s1 empty, it'll will run past the presignal.
If the s1 block isn't reserved, another train from the auxiliary track might take over that block instead,
leaving our train in the "never stop here" (super)block.

[ssilk]

Hm. I dunno.

And I think that is not that important, as in OpenTTD a, cause in Factorio we have much more space. That was always the biggest retention for me with OpenTTD (and other similar games): the sizes of the rails etc. weren't nearly halfway realistic. There is an obvious discrepancy between the length and the width of a track there.

Look at any map, search for the next train station and then look, how much space the needed rails, switches, waiting rails and so on need. Now compare that with OTTD: the difference is obviously immense, and the reason is, that those games really simplified much. A piece of rail in OTTD is how long? About 50-60 meters! In Factorio it is about 2. That is a scaling factor of 25-30! And how many other rails can I put in a space of 50 meter? 10 without much problem!

What I want to say: in reality those problems don't exists. Because of the space between all that. That makes such events as described really rare, and they are solved in a completly different way. For Factorio: that should be solvable with circuit logic. If you cabled your logic wrong: baaaammmm.

[t7119]

Hi Factorio's staff!
Video game development in Japan are looking yours.

"The questionnaire for the title you are looking forward to 2014"
http://www.4gamer.net/games/000/G000000 ... dex_5.html

Yoshinori Kitanose says,
This game is developed during the PC game. But very concept is interesting.
Players make the parts from resources such as iron and copper.
And make a huge automation planet!
The first is nothing in planet.
Eventually, will be filled in a lot of belt conveyor or a robot arm.
Steam locomotive galloping while spitting smoke tickles SF soul.

Who is Yoshinori Kitanose
http://ja.wikipedia.org/wiki/%E5%8C%97% ... 3%E7%AF%84
FinalFantasy series : Planer,Producer and Directer
Kingdom hears : Producer
Chrono Trigger : Directer and Scenario
Romancing Sa・Ga : Map Design

[Animar]

Oh man I'm already loving it to play arround with trains but all those upcomming features related to them getting me SOOOOOOOOOOOoooooo much exited. I mean Oil wagons, amasing Pre-Signals and maybe some more train logic like wait until full/empty etc... that just sounds so awsome that I want to play with it right now.

The only downside is that we have to wait for 0.12

[Jaridan]

Hm. I dunno.

And I think that is not that important, as in OpenTTD a, cause in Factorio we have much more space. That was always the biggest retention for me with OpenTTD (and other similar games): the sizes of the rails etc. weren't nearly halfway realistic. There is an obvious discrepancy between the length and the width of a track there.

Look at any map, search for the next train station and then look, how much space the needed rails, switches, waiting rails and so on need. Now compare that with OTTD: the difference is obviously immense, and the reason is, that those games really simplified much. A piece of rail in OTTD is how long? About 50-60 meters! In Factorio it is about 2. That is a scaling factor of 25-30! And how many other rails can I put in a space of 50 meter? 10 without much problem!

What I want to say: in reality those problems don't exists. Because of the space between all that. That makes such events as described really rare, and they are solved in a completly different way. For Factorio: that should be solvable with circuit logic. If you cabled your logic wrong: baaaammmm.

isn't this "cabled/circuited logic" just another word for enough options to handle trains and reservation & train logic(where you can do a lot with basically a normal signal,2block signal (for intersections) and platform choosing signals, simutrans logic/wording here , since i've played that since it's awakening more or less )?

[ssilk]

Maybe.

But it's a good point: I'm a bit sceptical, why everybody wants to have the train logic exactly the same as in those other games: a window with many tabs where I can choose under which circumstances the train should arrive and/or depart. And so on.

Much of that logic is only needed, if you have not so much space. In Factorio you can easily place 3 stations into one chunk. Or more. One for every case you have!
In Factorio it would be possible to let circuit logic choose the "right" station.
And so on.

I won't say, do it so, but I would like to see, if that really makes sense in a game, cause in reality it makes.

[mazetar]

I'd love to get more logic components into the railway system, and some of the components in OpenTTD would be nice to have.
I do however agree with whats said by ssilk above, it doesn't have to be the same implementation of the components.

I would love to get pre-signals to output their state to the circuit network, and having some signal(s) which accept input from the circuit network to turn on/off.
Making the logic network able to check the contents of a cargo wagon and output signal(s) to the circuit network could make for very interesting and complex systems.

And now if we'd only get the ability to cross under(or above) other tracks then I probably will probably loose my job due to Factorio <3

[Cloner]

...if we'd only get the ability to cross under(or above) other tracks ...

Yes .. underground train passage please !

[ible]

Underground train tunnels, why didn't I think of that.
It fits the current theme of belts going underground.

[-root]

Just whatever you do in .12, do not change the corners of belts. leave them as they are.

Hopefully they change it.
(It makes no sense to discuss this, especially not here.)

The question is, what they change it to.

If its some smoother, better version of what we have, sure, all for it.

But if they remove the speed difference from the outer and inner belt lines, I feel like it would detract from the game. Its a good example of complexity arising from something very simple and it adds a really, really nice wrinkle to the game when you have to design your belts in such a way that corners don't slow them down overly much.

[MF-]

But if they remove the speed difference from the outer and inner belt lines, I feel like it would detract from the game. Its a good example of complexity arising from something very simple and it adds a really, really nice wrinkle to the game when you have to design your belts in such a way that corners don't slow them down overly much.

For the impatient readers: This post calls for KEEPING the lane speed difference in corners.
I agree to the point and it would be unfortunate to have anyone misinterpret it.

[MF-]

Underground train tunnels, why didn't I think of that.
It fits the current theme of belts going underground.

You didn't because a train is far higher than some dust/rock lying on the ground.
Sinking it underground in just one tile is riddiculously unrealistic.
Combined with the max incline/slope trains could climb, it could take lost of to do it.

What do you think about the solution proposed on [Idea] Train-proof facility instead?

[kovarex]

Sinking it underground in just one tile is riddiculously unrealistic.

Who says it should be in one tile? The train to ground piece would certainly be several tiles long.

[hitzu]

Who says it should be in one tile? The train to ground piece would certainly be several tiles long.

Do you really want to add new spreadsheet of train-on-slope sprites? ʘ̆_ʖಠ

[kovarex]

Who says it should be in one tile? The train to ground piece would certainly be several tiles long.

Do you really want to add new spreadsheet of train-on-slope sprites? ʘ̆_ʖಠ

Well, it is one of the possibilities, as it would be only needed in 8 directions, and the train sprites could be moved into the ram-only memory. The other possibility would be to hide it completely.