Friday Facts #416 - Fluids 2.0

[mmmPI]

[sarcasm] The feel of superiority while telling a newb how this illogigal mess works will be gone. whan a letdown

Hahaha, i don't think after 971 hours you are considered "newb" , at this point if you still don't understand the fluid mechanic it's not because you are "new" but i think because you never bother attempting to learn.

[Makka77]

I have an oldish though well specced PC. Many hundreds of hours logged as we mostly have here, played in every way imaginable from skin-of-your-teeth deathworld to editor mode optimising train networks to the nth degree, multiple modded runs, and all the other funs we have available from this remarkable piece of software.

The only thing in this game that has ever made me frustrated with "the game" or "the devs" rather than logistics problems is the fluid system. Why isn't it working right, why isn't it working at all, why have I spent so long on this? Also, why has my UPS tanked? I hoped for a long time (especially after Earendel's hint) that the fluid system would be replaced with something less obtuse and lower rent, and welcome these changes.

@Raigard The idea of tweaks to address some of the concerns mentioned here sounds great, and you have communicated well (as we adore many at Wube for). If the properties and limitations of the pipes are communicated this well you're onto a winner!

*Whimpering in the corner until release day*

X

[mmmPI]

why isn't it working at all ?

Do you have an example where you don't manage to use the current fluid system ? you could post it in gameplay help, i'd be happy to help debug while the new system is not yet available, i found a way to make it work in some situations, i can make like nuclear power plant , and an oil refinery to show you.

[Makka77]

why isn't it working at all ?

Do you have an example where you don't manage to use the current fluid system ? you could post it in gameplay help, i'd be happy to help debug while the new system is not yet available, i found a way to make it work in some situations, i can make like nuclear power plant , and an oil refinery to show you.

Thank you kindly mmmPI, but that build has long since been torn down. As I remember it i had two rows of oil refineries with the output pipework down the middle, and there were odd cases where certain refineries couldn't output into the system at all, despite there being enough storage/cracking and consumption to use everything. I even tried out a flare stack mod to burn excess to no avail.

"Work out the kinks" one might say. But I had spent many hours debugging this system in editor to make one blueprint to rule them all! It worked flawlessly! But I don't know what order I put the pipes down. Certainly a different order from the bots. Big sad face. It made the game feel unfair.

[Sirad]

[sarcasm] The feel of superiority while telling a newb how this illogigal mess works will be gone. whan a letdown

Hahaha, i don't think after 971 hours you are considered "newb" , at this point if you still don't understand the fluid mechanic it's not because you are "new" but i think because you never bother attempting to learn.

Yap. i never bothered. I just once read about whats wrong and said (insert any curse) played the Game and took the approach of brute force. (overdesigning)
Best decision. soon its gone. Lifetime saved, Useless knowledge avoided. (no offense meant to the elitists here) and i will not start any new factory unless the new thing is out.

[Panzerknacker]

why isn't it working at all ?

Do you have an example where you don't manage to use the current fluid system ? you could post it in gameplay help, i'd be happy to help debug while the new system is not yet available, i found a way to make it work in some situations, i can make like nuclear power plant , and an oil refinery to show you.

Thank you kindly mmmPI, but that build has long since been torn down. As I remember it i had two rows of oil refineries with the output pipework down the middle, and there were odd cases where certain refineries couldn't output into the system at all, despite there being enough storage/cracking and consumption to use everything. I even tried out a flare stack mod to burn excess to no avail.

"Work out the kinks" one might say. But I had spent many hours debugging this system in editor to make one blueprint to rule them all! It worked flawlessly! But I don't know what order I put the pipes down. Certainly a different order from the bots. Big sad face. It made the game feel unfair.

Good that it didn't work because there is not supposed to be 'one blueprint to rule them all' xD

[XT-248]

You were mistaken. I stated your point about train didn't seem clear the first time i read it, it doesn't mean i am willing to enter a discussion about it.
It was just a feedback after attempting to answer the thing you said you needed help to understand to explain why i couldn't adress all the points.

Other player may also point at the difficulty using the previous system they encountered, similar to the post i quoted, which would adress your previous point that you stated you wanted help understanding.

Or be interested to have a dialogue with you about trains, since you are "covering again the same ground", i feel this is not new to me.

No, look at the post that someone took the quote out of. It was posted in April in a locked thread shortly after I posted (other people were making personal attacks that I had no part in).

Post by XT-248 » Mon Apr 29, 2024, which was locked a few hours later.

viewtopic.php?f=236&t=111775&p=609354#p609354

Someone made a 'reply' I never responded to before the conversation was locked (I know this because I recognize the quotation as part of a larger conversation about teaching new people how to properly handle fluid simulation from 1.1.x, along with other topics when I first saw the quotation). The 'reply' in that locked thread was off the mark. Yes, there is a solution, but what was missed in that conversation is how to 'teach' it to newcomers in-game (tutorials/Factoriopedia/scenarios/others). They first have to learn the 'why/how' the fluid simulation in a pipe junction works unpredictably before coming up with a solution.

I was not talking about how I 'need' to learn the fluid simulation, as that is something I learned many years ago when setting up a refinery/chemical plant complex to make products for a megabase.

I did point out that it was a weird way to quote someone and continue an old conversation.




Random strangers may put forward various points relating to the topic that I posted recently. I think it would be interesting to have a developer, if they feel like doing so, weigh the pros and cons of removing fluid simulation as pertaining to what I posted earlier, which is why I quoted them specifically and not just anybody on the forum with a strong opinion.

If someone felt that a topic of conversation is not 'new,' why respond to my 'not-so-new' points that I recently posted about? Bear in mind that Raiguard posted some new info about the difficulty of making a certain change a few days ago pertaining to 'undoing' a change. I wanted to know from a game development point of view why the new way of teaching players in-game mechanics (Factoriopedia and the new tooltip) would have been insufficient for teaching 1.1.x fluid simulation. A random stranger is unlikely to know the answer unless they work at WUBE.

If someone wants to continue the conversation about teaching new people how to properly utilize pipes, junctions, pumps, and continuing 1.1.x fluid simulation in Factcorio 2.0, take it to PM before someone decides to come in and axe the thread.


Back to the current conversation, I tried to make a subtle reference to a train fluid wagon and how it may lose some usability due to the simulation-less fluid in pipes, and someone missed the point I tried to make. No one is perfect.

[Tertius]

(overdesigning)
Best decision. soon its gone. Lifetime saved, Useless knowledge avoided. (no offense meant to the elitists here) and i will not start any new factory unless the new thing is out.

Hehe, that's a very pragmatic approach. I don't know if I'm counting as elitist, but I always want to know how everything is working with every detail. I spend hours to inspect how the game engine handles stuff, and how I can benefit from the peculiarities of the engine. And what I should avoid, if I see the engine isn't able to handle some other stuff not efficiently.
However, after I built all that knowledge and occasionally telling people about it, I'm always looking for a simple solution that avoids the weaknesses of the engine but doesn't explicitly exploit its peculiarities, because the more sophisticated and engine-dependent a setup is, the more fragile it is at the same time. And I wand universal, solid factories, not fragile ones.

So I'm looking forward to the fluid changes, because it will just deliver that simple solution I always seek, without the need to inspect and explore too deep. You nailed it with "useless knowledge".

And if the new system makes fluid stations and fluid wagons obsolete, so be it. We still have ore trains to build. If it's the most efficient solution to build a single long pipe web for crude oil across all our explored regions, so be it. That's something new, something we never had in the past. It's the middle ground between the electric power network and the train network, and that's good.

[mmmPI]

Thank you kindly mmmPI, but that build has long since been torn down.

I see you found out the reasons why the build didn't work initially and don't need my help It does highlight 2 potential causes/things to learn, throughput and junction order. ( supposing the refinery isn't stopped due to output full / improper cracking ratios which is a 3rd one, quite easier to deal with supposedly ).

It can take quite some time, but i also think it's well demonstrated by now, given how many players manage to make megabases or you answer that there was indeed way to make the system work despite the quirks even if it means scrapping some attempts in the process.

I think once it's understood that junction order matters in some cases it's possible to design so that it doesn't cause any problem in game. Albeit it may require looking up info on the forum, or doing in-depth test in editor to find out how to so in the first place

Yap. i never bothered. I just once read about whats wrong and said (insert any curse) played the Game and took the approach of brute force. (overdesigning)
Best decision. soon its gone. Lifetime saved, Useless knowledge avoided. (no offense meant to the elitists here) and i will not start any new factory unless the new thing is out.

That's not my playstyle ^^ , i have learned how to math out geometric series for the game factorio, it was much more fun than to do it in another context, which i can use for fluid mechanic, old system, or for trains, or the future system of fluid mechanic which require part of the same math to estimate the flow overtime. Lifetime knowledge aquired !

On the other hand i feel there is no particular knowledge require to use pumps and circuits that is becoming obsolete with the new system, to me it's like learning a langage, once you are fluent with it, you can write a book, a novel, or a a tweet.

I'm not saying it is necessary to enjoy the game to be willing to read about the fluid simulation, but i think i would have been bored not learning things that are useful for me beyond the videogame world after some of the thousand of hour i played. To me useless knowledge would refer to things like skill trees in MMO or maps from imaginary places that i still remember from playing other game. I'm just saying this because i think it is always possible to learn new things, and sometimes it's only after you learn about them that you realize how useful this newly acquired knowledge is ! I do not think of myself as an elitist, i took no offense, i encourage everyone to try and learn new things, i think elitism is considering some people can't or aren't going to learn and should be treated differently. I try to push everyone to learn things that i find valuable after i learn them !

If someone wants to continue the conversation about teaching new people how to properly utilize pipes, junctions, pumps, and continuing 1.1.x fluid simulation in Factcorio 2.0, take it to PM before someone decides to come in and axe the thread.

I agree, i'm going to not answer anything from you for that reason.

[XT-248]

Best decision. soon its gone. Lifetime saved, Useless knowledge avoided. (no offense meant to the elitists here) and i will not start any new factory unless the new thing is out.

It may not end up as you think since there will be a new set of useless knowledge with the simulation-less fluid in place.

Hehe, that's a very pragmatic approach. I don't know if I'm counting as elitist, but I always want to know how everything is working with every detail.

*snipped*

And if the new system makes fluid stations and fluid wagons obsolete, so be it. We still have ore trains to build. If it's the most efficient solution to build a single long pipe web for crude oil across all our explored regions, so be it. That's something new, something we never had in the past. It's the middle ground between the electric power network and the train network, and that's good.

I don't consider myself an elitist by anyone's standards. I also like to sit down and learn how mechanics work under the hood.


I disagree that a simulation-less pipe is a middle ground between an electric power network and the train network.

The pipe has a physical presence just as much as the electric network does and moves either fluid or energy, respectively, in a simulation-less manner from source to destination. The train network still has a physical presence and a simulation aspect, in the form of train engines and ore/fluid wagons, and needs to go from A to B point. There is very little overlap between a train network and simulation-less pipes.


I also disagree that designing a pipe network to carry everything is 'something new' since we could already do so with simulation fluid. The only thing that will change is the number of fluid entities that can move fluid that we need or don't need.

If someone wants to continue the conversation about teaching new people how to properly utilize pipes, junctions, pumps, and continuing 1.1.x fluid simulation in Factcorio 2.0, take it to PM before someone decides to come in and axe the thread.

I agree, i'm going to not answer anything from you for that reason.

I am fine with that.

Let's all move on.

[mmmPI]

hopefully the new system will see less people complaining that the logic is absolutly flawed because it's easier to use
and that everyone will be able to finally use pumps to control where the fluid goes

[GTexperience]

Hey Everyone,

I thought I might just post some concepts on how the fluid system could be improved even further. It's not perfect and is just a proof of concept for one input and one output with a pipe length of n. What if we could still allow for fluid falloff within the pipe systems? It would prevent instant fluid transfer, as well as, prevent pumps from becoming mostly useless (useless except for trains and circuits). As when consumption and production of an entire system are balanced, pumps are completely redundant. This is just a concept of how we could add a linear falloff to the pipes while still treating them as one group. Hence, it still avoids the problems presented by the previous system that allowed for backflow when there shouldn't be any. Per group, the linear falloff would only have to be calculated once per frame as it is consistent for each pipe. The linear falloff would be calculated based on the input and output ratio, as well as the average amount of liquid in the pipes.

Please note, this is just a concept that might not even work and surely has its own drawbacks! It's just an idea I had in my head that I wanted to share regardless.

Code: Select all

// Function to calculate the skew angle using weighted linear interpolation
double calculateSkewAngle(double inputRate, double outputRate, double averageContent) {
    double angle = 0.0;

    // Calculate the content ratio (0 for empty, 1 for full)
    double contentRatio = averageContent / MAX_UNITS;

    // Calculate the input/output rate ratio (clamped between 0 and 1)
    double inputOutputRatio = std::min(inputRate / (outputRate + 1e-9), 1.0);

    // Calculate the skew due to content (0 degrees when full, 90 degrees when empty)
    double contentSkew = 90.0 * (1.0 - contentRatio);

    // Calculate the balanced skew angle based on a 100 by N right scalene triangle
    double balancedSkew = std::atan(100.0 / N) * 180.0 / M_PI;

    // Calculate the skew due to input/output rate (balancedSkew degrees when balanced, 90 when input >> output, 0 when output >> input)
    double rateSkew = balancedSkew + (60.0 - balancedSkew) * (inputOutputRatio - 0.5);

    // Combine the skews
    angle = (contentSkew * rateSkew);

    return angle; // falloff angle
}

[mmmPI]

It's not perfect and is just a proof of concept for one input and one output with a pipe length of n.

How are we supposed to use the skew angle ? apart from calculating it , i don't undertand how it should be used in relation with fluids quantity over time

[XT-248]

hopefully the new system will see less people complaining that the logic is absolutly flawed because it's easier to use
and that everyone will be able to finally use pumps to control where the fluid goes

I wonder why someone would assume that people don't use pumps now? Given that there are untold numbers of publicly available screenshots of pumps being used by veterans and newcomers alike.

Hey Everyone,

I thought I might just post some concepts on how the fluid system could be improved even further. It's not perfect and is just a proof of concept for one input and one output with a pipe length of n. What if we could still allow for fluid falloff within the pipe systems? It would prevent instant fluid transfer, as well as, prevent pumps from becoming mostly useless (useless except for trains and circuits). As when consumption and production of an entire system are balanced, pumps are completely redundant. This is just a concept of how we could add a linear falloff to the pipes while still treating them as one group. Hence, it still avoids the problems presented by the previous system that allowed for backflow when there shouldn't be any. Per group, the linear falloff would only have to be calculated once per frame as it is consistent for each pipe. The linear falloff would be calculated based on the input and output ratio, as well as the average amount of liquid in the pipes.

Please note, this is just a concept that might not even work and surely has its own drawbacks! It's just an idea I had in my head that I wanted to share regardless.

*snipped code*

I will say one thing about the code: Double is well-known to be memory-heavy, slower, and prone to loss of precision. Perhaps consider using Integer or some variants instead. IE: What we know now as 50 fluid units is Integer 5,000 in the background. Integer / 100 = X.XX for displaying only.


Otherwise, I think there is room to improve the concept some more.

[Sirad]

It may not end up as you think since there will be a new set of useless knowledge with the simulation-less fluid in place.

The new system is intuitive. Thats enough for me. Build your design. it works or not. But if it works or not, does not hang upon what order you build your pipes or what kind of junction its in. I dont need to state again what kind of knowledge about Pipes i consider useles nor ?

[mmmPI]

hopefully the new system will see less people complaining that the logic is absolutly flawed because it's easier to use
and that everyone will be able to finally use pumps to control where the fluid goes

I wonder why someone would assume that people don't use pumps now? Given that there are untold numbers of publicly available screenshots of pumps being used by veterans and newcomers alike.

I am basing this on quotation like this one where you clearly stated you can't use pump to control the fluid :

viewtopic.php?p=609354#p609354

Even increasing the number of pumps and fluid throughput continues to demonstrate that I cannot control where the fluid goes in the same way a splitter can control where items go with priorities and filters.

I know there are untold numbers of available screenshots of pumps being properly used by veterans and newcomers alike. But occasionnaly there was a player that couldn't manage to do so and blame the game, like you and some others did in the post i quoted. I hope this is going to change with the system now made easier.

[XT-248]

It may not end up as you think since there will be a new set of useless knowledge with the simulation-less fluid in place.

The new system is intuitive. Thats enough for me. Build your design. it works or not. But if it works or not, does not hang upon what order you build your pipes or what kind of junction its in. I dont need to state again what kind of knowledge about Pipes i consider useles nor ?

I am not saying that the new system is not intuitive.


I am saying that a new set of useless knowledge will exist without fluid simulation.

[XT-248]

hopefully the new system will see less people complaining that the logic is absolutly flawed because it's easier to use
and that everyone will be able to finally use pumps to control where the fluid goes

I wonder why someone would assume that people don't use pumps now? Given that there are untold numbers of publicly available screenshots of pumps being used by veterans and newcomers alike.

I am basing this on quotation like this one where you clearly stated you can't use pump to control the fluid :

viewtopic.php?p=609354#p609354

Even increasing the number of pumps and fluid throughput continues to demonstrate that I cannot control where the fluid goes in the same way a splitter can control where items go with priorities and filters.

I know there are untold numbers of available screenshots of pumps being properly used by veterans and newcomers alike. But occasionnaly there was a player that couldn't manage to do so and blame the game, like you and some others did in the post i quoted. I hope this is going to change with the system now made easier.

#Picard_Face_Palm_Meme

I demonstrated knowledge of how to break the fluid simulation in a manner that makes it impossible to control which machines on a branch of a junction receive fluid first to last in that thread.


Furthermore, if one reads the entire post from beginning to end. One will find this sentence, which further drives the point I was trying to arrive at.

Since the point is to demonstrate that pipe junctions are black boxes that the players have no discrete control over beyond picking up and placing pipes without knowing what they would do afterward.

The problem is how to teach players how to control the fluid in an intuitive way in the game itself.


Think about it: how could I create a demonstration of fluid simulation that doesn't work intuitively if I didn't know how the fluid simulation works?

If I were seeking a solution to the junction issue with fluid simulation, I would not debate the merit of keeping the fluid simulation by teaching it in-game to players better.

[mmmPI]

Even increasing the number of pumps and fluid throughput continues to demonstrate that I cannot control where the fluid goes in the same way a splitter can control where items go with priorities and filters.

The problem is how to teach players how to control the fluid in an intuitive way in the game itself.

I agree and hope the new system will be easier than the old one for everyone to control with pumps, good debate sir.

[picklock]

Another example of how game mechanics are optimized here, even if it sometimes takes a little longer and in this case was triggered by the expansion. Can't wait to test this new feature in the game myself.