Friday Facts #133 - The train struggle

[taiiat]

a more practical implementation may be.... to allow Inserters to get the stack size bonus they get for picking up from machines and Et Cetera, to picking things up from belts.
which would let Inserters 'scoop' items from a belt. no new type needed at all, just probably a Research. but limited to fast/long/smart Inserters i'd imagine.

if Loaders or Heavy Inserters, Loaders actually probably less disruptive than Heavy Inserters in terms of complete obsolescence...ness.

[RobertTerwilliger]

I'd rather see a gap in horizontal and no gap in vertical train stations - that stretching doesn't look right.
I know you guys are trying best, but I think the best solution is just fix wagons to grid, so all wagons have similar inserter setup, and leave us with the fact we have to have saparate H and V station design - it'll be the most good-looking decision IMO.

[The Phoenixian]

No mention of a fix for crashing on autosaving

Bug fixes are rarely mentioned during the Factorio Friday Facts. The few times they do are usually very weird bugs or discussions that illustrate the process of bugfixing.

For what can be done, If you have the spare time and energy needed to give the devs more info on your problem, I advise you to do so. There's no obligation to do so, but it would make their work easier and help fix the problem faster.

a more practical implementation may be.... to allow Inserters to get the stack size bonus they get for picking up from machines and Et Cetera, to picking things up from belts.
which would let Inserters 'scoop' items from a belt. no new type needed at all, just probably a Research. but limited to fast/long/smart Inserters i'd imagine.

if Loaders or Heavy Inserters, Loaders actually probably less disruptive than Heavy Inserters in terms of complete obsolescence...ness.

I think this depends a lot on precisely how heavy inserters end up working. If they're significantly more expensive, then even being better in every way may not be worth the expense in low throughput scenarios where inserters are not the limiting factor. Even beyond that, the mechanics of picking up and setting down items may mean that blue inserters are better on occasion, as is already the case with other types.

For instance, even burner inserters see some use in the mid game (and late game if you don't switch to solar or use steam backup power) simply due to not being vulnerable to electrical shortages. Likewise the yellow inserter picks things up slowly, which can be useful for not draining a belt too fast. The very thing that makes them inferior in most contexts grants them a lasting niche in certain situations.

If a heavy inserter takes more time to pick up or even just set down objects, that may end up making it slower even if it has the same turn speed.

Likewise, there are a few cases where the stack size bonus is a pain, rather than a benefit: The stack size bonus makes it difficult to split building outputs when producing more than 2 items per run. (See, optimal green circuits.)

[Ironhand]

Maybe the trains could stretch a little and the gap between them could stretch a little too. There's gotta be a sweet spot where both distortions are less noticeable.

[LazyLoneLion]

Even Factorio itself has run into this with a few aspects of it's development, like multiplayer: The issues with descyning required a huge amount of bug fixing and that ate up massive amounts of dev time.

But it did became better, didn't it?

[AssaultRaven]

Even Factorio itself has run into this with a few aspects of it's development, like multiplayer: The issues with descyning required a huge amount of bug fixing and that ate up massive amounts of dev time.

But it did became better, didn't it?

Only if you play multiplayer.

[Earendel]

I think that the train fix is important for a number of reasons, not least blueprint rotating.

In the distortion example it looked like linear interpolation, but I think it would be best to use sin or cos to get the distortion factor.

A bit of distortion in conjunction with a bit of extra spacing would probably do the job.


The projection issues are a problem for a lot of things, not just train graphics. It's a problem for all movement speeds, anything with a circular range (turrets) distribution of trees, etc. I think that having a grid scale that more closely matches the projection distortion would be worth doing despite the graphic headache. I suggest keeping the cell width the same, and reducing the height by about 20-25%. For most graphics, just use the current graphics and they will just overlap slightly, that shouldn't cause too many problems. For tillable graphics, just scale them along the Y axis so that they fit the new grid. That would apply to the belt, tracks, pipes, and terrain. Doing it that way would also prevent most mod graphics from breaking. Pipes would need to be re-rendered to the new size to keep them looking circular, the the distorted versions shouldn't look too bad.

[The Phoenixian]

Even Factorio itself has run into this with a few aspects of it's development, like multiplayer: The issues with descyning required a huge amount of bug fixing and that ate up massive amounts of dev time.

But it did became better, didn't it?

It depends on whether or not you consider it worth the cost. I think that Multiplayer was worth it, even with the fact that it's first release was in 0.11 October of 2014, 0.12 took until July of 2015 and more work on multiplayer still remains to be done. (And we also had all the work on replays in 0.10 and earlier versions as a precursor to and foundation for multiplayer.)

For a reference, It's worth considering that had the devs not added multiplayer, the added time might not have merely let them work on the space platform, but we might even have it at this point.

Or if that hadn't been decided as a focus, any of several things on the roadmap. (Not all mind you, but a lot of the 0.13 features might already be present.)

[KatherineOfSky]

The Heavy Inserter looks very interesting. I'll be anxious to see how it stacks up against the Fast Inserter in actual game play.

Glad to see that work is being done on the train system -- while the current situation is not bothersome to me, it will be a good quality-of-life improvement, especially for rotation of blueprints.

I like the idea of having gaps between each car for measuring/spacing. The current incarnation is less-realistic, and connection gaps where the wagons cannot be loaded would add to the game.

[VDOgamez]

While I like the idea of having the train lengths rotation invariant, the current method you're looking at just looks terrible. It might be a better idea to just accept that vertical stations are shorter, and simply work to fix the cars to integer lengths, unless you can find a much better-looking solution.

The heavy inserters make me think of those clamshell buckets you sometimes see used on cranes for picking up big piles of stuff, which appeals to me. How would those interact with the inserter stack size bonus, though?

[vliedel]

How about you make a new type of train wagon. For example, a wagon with a container, which can be taken off in a short time (an empty one will be put on the wagon again). Then, while the train is gone, the goods are slowly taken out of the container.

[Albert]

the current method you're looking at just looks terrible. It might be a better idea to just accept that vertical stations are shorter, and simply work to fix the cars to integer lengths, unless you can find a much better-looking solution.

The screenshots from the game showing the new plan for the train are terrible. Agreed.
However, I wouldn't worry too much about it, they are just showing that we can reduce drastically the amount of distortion. Once we finish with this subject I'm pretty convinced that the new trains will look very reasonable, and this painful geometric problem will be visible just for the most expert eyes, only. Hopefully.

This is not an invitation to show in public how expert eyes you have once we have the final art for the trains. If you know the issue probably you will notice it, but in here we're smoothing the problem as much as possible.

Keep in mind that this problem, squared grid VS visual perspective, is chasing us since the very beginning.
Concerning this issue, trains are the most problematic entity at the moment, because they move along different axis, X and Y, and the engine doesn't give a damn about Z, but the perspective does.
So the height is always an issue for the gameplay. The engine thinks that Factorio is a top down game, but the game doesn't look like that.

If the entity moves from axis to axis, but is also attached to the grid, the problem is bigger. Imagine the pain behind the transports belts, which are also tiled and animated. Most of the time we need to do hard tricks in order to fix this, not to mention to make it fit with a proper visibility for the gameplay and nice looking.

If I would start Factorio from scratch again, maybe I'd use a different perspective, or maybe not, or a different grid. But the use of the distorted grid, the wide rectangles, I'm afraid that this would cause a lot of problems also, just different ones. This subject requires a deep study and time goes by...

cheers!

[Neotix]

But why you want to fix something that is not broken in the first place? Only to be able to rotate blueprint? Really?
After that changes you will take away some solutions for shorter stations and for shorter vertical waiting tracks. For me that changes are very bad idea.
Only thing that need fixing is train alignment to the grid on stations. There is even mod that trying to fix that but it's workaround.

[AssaultRaven]

I realize that the other reason why this is a problem with trains specifically is that they have such a large length/width ratio compared to anything else in the game. Instead of distorting them, what about just making train objects more square? Wouldn't that let you distort them without it being noticeable, much like how many people don't even notice that conveyor belts aren't square without it being pointed out?

Like this:





Also mostly solves the difference in capacity of cargo wagons vs chests, and feels to me like it would more closely match the scale of the rest of the game, also.

[AssaultRaven]

Actually, I wonder if the height wouldn't still pose a problem with those.

Well, in any case, this isn't Railroad Tycoon. There's no reason why the trains have to look like they do now, but should instead by redone first and foremost with regard to the art-style and projection. "Wooden chest with wheels" might be going too far, but there's no problem if they look closer to that if it means avoiding the horizontal/vertical problem entirely.

[onebit]

I'd rather have fluid wagons than train length fix.

[Ranakastrasz]

The animation of the train rotation issue makes it clear that it will be visible, so cannot be ignored, which is actualy proof, deraling aruments on that part of the subject.

The heavy inserter, honestly, is something I think all inserters should do anyway, but I see why you might want to restrict it.
At minimum, if/when it is released, I will make a mod that makes all inserters have that functionaility, assuming it is just a flag instead of a different type of object.

[againey]

But why you want to fix something that is not broken in the first place? Only to be able to rotate blueprint? Really?
After that changes you will take away some solutions for shorter stations and for shorter vertical waiting tracks. For me that changes are very bad idea.

Different play styles will inevitably lead to different conclusions about the significance of this stuff. You consider it not broken, and in fact supplying additional game options. I get that. But with my play style, I feel forced to never build vertical stations, because I want to maximize inserters on each car. This leads to some painfully awkward layouts in some cases, which really irritates me. So yeah, it feels broken to me.

Besides, it's like one of those situation where once I've see it, I can't unsee it, and it always irritates me when a game's mechanics aren't self-consistent. In this case, it feels like the game world can't decide if it's top-down or an angled perspective, as glex described. Clearly, that can't be solved perfectly without redoing a massive amount of the game. But my personal preference would very much be to keep all the inconsistency out of the actual mechanics of the game, keeping it truly top-down and square-based at its core. Let the visuals take the full brunt of managing the inconsistency as best as they can. Because it doesn't actually affect how I perceive the game; only the mechanics do. But I know not everyone prioritizes mechanics over aesthetics like I do, hence the debate. So I'm not saying you're wrong in your preference; just suggesting that I'm not wrong either, even though we want different things.

[Koub]

The animation of the train rotation issue makes it clear that it will be visible.

Depending on the way you show things, they can be obvious or not :
https://www.youtube.com/watch?v=A4QcyW-qTUg

I'm sure a train just doing its train things and being stretched or compressed according to its orientation would remain unnoticeable without tricks explicitely meant to help actually see it happen.

[RobertTerwilliger]

Guys, it's probably time to VOTE to help devs with this issue.
Who knows, maybe they're trying to make unnecessary job now, and there's an easier solution that will be warmly accepted, or we will encourage them to make it exactly in the way they've already started it