Fluid Optimization not fixed for release?

[konnan44]

Is the game still in an Alpha/Beta stage? Why was it rushed to release. One of the most frustrating things for me has been how fluids work, and it was never fixed. How is it possible to have two pieces of pipe connected to each other, and one pipe is 50% full, the other is 20% full. Can I please see a real life demonstration of this?

[DaveMcW]

https://www.factorio.com/blog/post/fff-271

We are tackling the fluid changes in 3 stages:

1. Move all the fluid logic into a separate system.
2. Merge straight sections of pipes into segments.
3. Tweak the fluid flow logic, which will not be an optimisation, but a gameplay mechanics improvement (FFF-260).

In the end, #1 was implemented, #2 and #3 were cancelled.

[5thHorseman]

I'm not happy about it either but to say Factorio was rushed is pretty funny.

Also most of the best parts of the game don't have any real parallels to reality. It stands to reason the bad ones don't either.

[Serenity]

Optimization isn't the right word for this. They optimized the fluid system a great deal

The alternative algorithm however didn't work out. They tried, but there were unexpected issues with oscillations building up

[Koub]

It's not excluded they might come back to it later. Despite the game being out of early access, there are still tasks they deem worthwhile working on.

[Hannu]

I'm not happy about it either but to say Factorio was rushed is pretty funny.

Hey, it is able to handle only 100000 entities with upper middle class desktop PC and have some inconsistency of fourth decimal in fluid flow. Score is 15/100. Real game should be able to handle billions of entities and simulate fluid flow at quantum mechanical level. However, I would not hold my breath or leave Factorio without playing during wait.

[quinor]

From what I know, the fluid flow logic redesign had encountered a few really exotic nad difficult problems that the guys were unable to solve. It was in my opinion a difficult decision to make to cancel that, but I hope they'll come back to it in the future and have another take on it.

[Yoyobuae]

I do wonder, what kind of factory this guy is building that the occasional fluid system wonkyness completely ruins his experience?

[netmand]

Like it or not: You had to have had followed all of the FFF blog and related posts to get a good explanation of the current fluid system.

Asserting any definition of real fluid dynamics upon the fictional universe in which Factorio is based, is in error. It is fair to want to know what that definition is; which is part of the game. And to me one of its more endearing aspects.

However, we can still report bugs and odd behavior. It would be a lot more helpful if the report was more than, "one pipe is 50% full and the other is 20% full".

[Aru]

If there were unexpected issues, they did it wrong.

To this day, any design I make with fluids is constantly foiled by the fluid physics. I suggested a solution 3 years ago. They haven't implemented a fix because they think it's too much work to write the known solution into code.

[eradicator]

If there were unexpected issues, they did it wrong.

To this day, any design I make with fluids is constantly foiled by the fluid physics. I suggested a solution 3 years ago. They haven't implemented a fix because they think it's too much work to write the known solution into code.

Have you tried requesting source access and implementing it yourself? I bet there's money to make there. And a lot of people to make happy (me included).

I think the main problem with fluids is that everything else has super precise numbers on it. But fluid flow is really difficult to calculate even if one does know all the variables, build orders, chunk boundaries etc. I can see how that might frustrate people.

[Aru]

By the way, the OP is about the fluid mechanics, but the title is fluid optimization. Optimization means making the game run faster, mechanics is changing the results, it's the wrong word. Though there is an indirect connection, in that anything that works also has to be fast enough for real time calculation in a game, on top of everything else the game has to calculate.

[Koub]

If there were unexpected issues, they did it wrong.

To this day, any design I make with fluids is constantly foiled by the fluid physics. I suggested a solution 3 years ago. They haven't implemented a fix because they think it's too much work to write the known solution into code.

Unless you are a secret member of the dev team, you're in the Fundamental attribution error realm.
You can't know what they have tried, for how long they have tried, what has failed, and why it failed.

Maybe you'll understand more reading this :
https://www.factorio.com/blog/post/fff-260
https://www.factorio.com/blog/post/fff-271
https://www.factorio.com/blog/post/fff-274

and these posts from Dominik :
viewtopic.php?p=457135#p457135
viewtopic.php?p=450523#p450523 and all the following posts by Dominik

As you can see, the fluid system overhaul has been worked on for roughly one year, before it was dropped out of the 1.0 scope (which doesn't mean it will never make it to the game, just that the game would get out of early access with a perfectible fluid system).

[Aru]

I read the blog posts before.

[t-lor]

If there were unexpected issues, they did it wrong.

To this day, any design I make with fluids is constantly foiled by the fluid physics. I suggested a solution 3 years ago. They haven't implemented a fix because they think it's too much work to write the known solution into code.

People have 1 hour done speedruns. make 10k megabases. and did everything in between while burning down the world with a bazillion flamethrowers.
Im not saying it fluids are perfect.

But if you cant get any design to work, i have a feeling thats saying more about your design then the game :0

[5thHorseman]

If there were unexpected issues, they did it wrong.

nobugs_jesus.jpg (125.5 KiB) Viewed 7598 times

[Aru]

But if you cant get any design to work

That's not what I said. You three are deliberately misinterpreting.

[Deadlock989]

That's not what I said. You three are deliberately misinterpreting.

No, I think they understand you pretty well.

[Hannu]

To this day, any design I make with fluids is constantly foiled by the fluid physics.

Is there any practical example in which there are problems? I do not mean experimental setups intend to achieve some exact theoretical throughput but practical production plant which do not work, let's say at 95 % of theoretical throughput. I have made hundreds of fluid systems (including large Angel's oil refineries with tens of liquids and complex combinator control) and never encountered throughput problems due to inconsistent fluid dynamics.

[tk0421]

I only have one complaint about fluids: I miss 3-type configurable liquid wagons.... (yes I know there's a mod for it)