Friday Facts #392 - Parametrised blueprints

[Weeno]

This is totaly awesome, i can't wait!
Now if we could reuse blueprints in other blueprints and have them update automatically, there would be nothing more to wish for

[DanGio]

Thanks for sharing this ! IMO, it's a really nice to have feature for cleaning my train book. That's about it. I mean, I already have several robomalls blueprint that carry a special feature (like iron stick assembler to avoid carrying those by bots) so they're not replaceable by a generic one.

Other than that, well, we'll see what the extension holds for us, and also, we'll see what genius ideas for dependant parameters & numbers players will come up to.

edit : spoiler : I fear there won't be many unless you implement the 1000 features we now want on top of the presented ones

[Serenity]

Something that would also be really useful is a generic belt / underground / splitter that gets assigned its color when you put it down. It would clean up balancer books for example. And it's a way to introduce this feature earlier, or for people who aren't so heavily into complex train setups.

But then I guess you'd have to think about what else this can be used for. And all other tiered machines like assembly machines, furnances and inserters (at least normal, fast and stack) also have to be included.

[mmmPI]

Brilliant statement that blueprints kill creativity in general.
Allow me to show you one extreme example to that theory. Could be the future, who knows.
I mentioned that as watching how new players use the feature will provide some insights into your theory.

The video is interesting, it could lead to situations where it ruins the fun for a player because they can just ask the AI what's next ?, how to do ? and then the game becomes obey the AI, i suppose that's one of the risk you refers when you make the link between blueprint harming creativity 'in general' and the AI playing minecraft for the player.

I think though it is doing lots of efforts to mimicks things it doesn't really do, like giving itself tasks , it is relying on humans understanding/experience of the environment 'manually' translated in a sort of graph where the AI can navigate "logically", and then make a corresponding output answer ( a task ) to any input ( description of where character is and surrounding).Then there is a step where the "task" is changed from concept to actual code to run in the game that's some pretty impressive code, but that's not the AI that came up with it. It's some humans who made up what is a "task" then the machine is attempting to make code that would "pass" the human evaluation of "is task accomplished". Either by collecting knowledge from aggregated human behavior to know what is a minecraft what task exist and so on, or that part when it's described how they teach the ai to debug its own function, this is different from other AI than learn the rules of games by themselves in comparaison, it's not the same way of achieving result. The kind of AI that learn how to play the game by themselves are like those that play chess or go. They can't explain to human their strategy because they are not using human understanding/experience of the environment. The "blindness" of the "fancy chatbot" is cleverly dealt with, but that is still a fundamental difference considering what/where/who is "the smart".

For factorio blueprint, i think "balancer" is good concept to illustrate what would be and "AI ruining creativity" it would be the AI telling you to make a balancer and how to make one, but that is not the same as the AI that would learn to play factorio, because such AI would not use archaic things like "balancer" , it would build a factory that would not be imbalanced or maybe only using direct insertion as it could manage to find how to most efficiently build things in a way that maybe will result in a monolitic mess of things that contain just the smallest amount of component to achieve the task it gave itself, and such task may be impossible to formulate fo human understanding, because maybe it has to do with balancing the inner state of the machine based on its own understanding of the environment.

You don't need an AI to ruin creativity, a blueprint portal would do the same as a "fancy chatbot" AI, and a go-player AI may not yield useful things for humans.

But a blueprint portal , similar to fancy chatbot is only fed by humans, considering the individual player it can ruin its own experience, but it can also be that the player try to participate in the portal, posting their own blueprint, receiving feedback on it, and improving their skills or solving problems, getting unstuck or learning things through it. The portal if considered like a fancy chatbot is the sum of the knowledge from the players, but it wouldn't add things on its own, it wouldn't know what is a balancer, or what is a train station , and how those are like "task" that the player may need as a single conceptual entity. ( like the blueprint portal category would have to be made by player, maybe with a tag system so it's aggregating human understanding). This has also a potential to enhance creativity, let say "train junctions", it create some sort of emulation where the ideas and tricks of different players are gathered in one place, and can be combined or reused to create designs better than any individual player would have gotten on its own. It can also create some "fill in the gap" mentality, you see some nice balancer blueprint for 2=>4 and 2=>8 but you need a 2=>6, and so you add it to the list once you've done it, or you post your attempt in the portal, and the list extend, and maybe someone will do a 3=>8, and what nots. It is also the case for combinator, someone post a memory cell design on the portal, and someone post a different one, and another one and so on, and maybe it become a mini game in itself.

For the future i don't know, for know it seem you can't ship a game with a smart assistant in it at release day, as it require being fed by human experience, but it could open potential for very good "tutorial" where the game understand what you don't understand, and how to teach you only that after looking at you playing for a few moment instead of asking you wether or not you need tutorial. If you consider it an AI and not a mod portal, it could be maybe in the future a comon thing that people fed one together and compare their method of feeding an AI as part of a game. I don't think it has to ruin creativity, i think it can be a enhancing tool if integrated properly.

Recent video of his showing some of what that may look like: https://www.youtube.com/watch?v=HzpUQZIr15g

Oh very good video too ! It make me feel better about my difficulties using the mod, it shows that it really requires lots of effort and talent to make to good use of it for a task that can be stated in a few words , i think it also kind of show a glimpse at how difficult it is to understand a build that was made by someone else, i imagine an ai that learn to play factorio by itself would create contraptions whose complexity would be much worse regarding how difficult it is to understand what was built and how it works.

It also shows i think how a "blueprint portal" may not kill creativity, imagine if the many single parts of logic where pieces available to all player ? like shift register, memory cell, timer, loop, signal sorter and things. That's pretty much what i'm using as blueprints, it think it's not against your idea that in order to use blueprint one need to make its own smelter first, to understand how it works, i really share that feeling, i think for combinators it's impossible to combine the different parts if you don't understand how each works individually anyway but having a library of those i think is very valuable, exploring each one of them is "a lot of content" and "opportunity to learn". I was a bit sad to see the impact of the blueprint introduction on multiplayer games because it somehow lead to a phase where more players where mindlessly flling in the blueprints the host pasted, and if that was a bad blueprint or disproportionnate or whatever mistake the host did, it would lead to no discussion or creativity and counpound overtime but now if you got a portal you could just refers to the pages or discuss about the choice of the blueprint for the particular game, it could/can restore a bit of that teamwork. I think it can also illustrate situations when an AI tasked/programmed by humans to assist oher human would be different conceptually than an AI that tries to beat the game and how there is room for humans to express their creativity in the making and continuous improvement of an automated assistant / blueprint portal.

I try to be optimistic here okay, how to make the best of the future

If there was a tutorial, as mentioned in that nVidia demo, that'd change things for newer players. Also there's not a tool like Helmod built into the game, likely for a reason similar to your theory. If i'm playing vanilla, I'm more likely to use blueprints, but if I could install helmod, I'm more likely to not use any blueprints.

I do believe adding a blueprint hub, similar to the mod hub we have in game, would cause people to play the game more. For some of the reasons mentioned above of my use case.

For helmod/ factory planner or other tools that help to math ratios like rate calculator, i think they can be used as tool to make better blueprint 'faster', thinking the understanding of "why" you need ratios in the first place is the most important, but then you may want to skip on calculating the impact of moving the beacon by 1 tile, everytime you fiddle around when making a blueprint for a production lane. I play often modded games with new items and productions lanes, but the blueprints for train station and combinators stays the same, only some parameters changes , production lane require "game blueprint" which do not carry over using different mods.

Expecting players to use the tools to maximize their enjoyment is not proper assumption though i think. That's the paradox of considering every new player know which blueprint to learn from and when and which to keep for later, or the difference if you make a shortcut that makes people's shoes dirty, and expect only people who already have dirty shoes to use it and realize there also exist people who just care about time and would take the shortcut even with clean shoes and blame you for making the dirty shortcut existing/available. Imagine getting negative feedback from player that couldn't find the blueprints needed to complete the game and that much time was spent in frustrating search for the "melanger part 2" whatever that may be and all you have to answer is " no but you have to make it yourself that's the point of the game"....

It's hard to stay optimistic !

[mmmPI]

Something that would also be really useful is a generic belt / underground / splitter that gets assigned its color when you put it down. It would clean up balancer books for example.

Some balancer with blue underground belt can't be made red or yellow though i tried to clean up to only keep the blue book and use the downgrade planner, but it didn't work out that way. i kept yellow and blue and use upgrade planner on the yellow for red.

Though thinking of this now, parameter for blueprint would be of good use for sushi belts !

[Xuplus]

This is awesome kovarex, many thanks!

Would it be possible to define limits to some of the parameters?
So in the "Parameterised build" dialog, instead of letting you input any value, you are limited to a set of options, or a range of numbers.

[Serenity]

Some balancer with blue underground belt can't be made red or yellow though

Yeah, you have to be a bit careful, but they generally work upwards, if not downwards. Blueprints that are universal could be completely generic, and others can use a specific color for a belt piece when needed

Another simple example would be furnaces where you can just replace stone with steel and yellow with red belt.

[Krissy]

What do I think? I think I don't want to play this game again until this is playable. Please let us play test it for you!

[Annihilannic]

It's "parameterisation" (UK) or "parameterization" (US)... pretty sure it is "parametrisation" nowhere?

[mmmPI]

It's "parameterisation" (UK) or "parameterization" (US)... pretty sure it is "parametrisation" nowhere?

That's how it's spelled in french, but apparently it's also a valid spelling in english , or at least in mathematics.

In mathematics, and more specifically in geometry, parametrization (or parameterization; also parameterisation, parametrisation) is the process of finding parametric equations of a curve, a surface, or, more generally, a manifold or a variety, defined by an implicit equation

https://en.wikipedia.org/wiki/Parametri ... (geometry)

[FasterJump]

I've been wanting this for years.
Would it also include the ability to modify modules?
For example, lets say I have an old blueprint that uses assembly machines 1's. I then use an upgrade planner to upgrade the assemblers in that blueprint to tier 3's. Now, there are more module slots available. Would I be able to fill in those module slots (or even better, parameterize) those module slots?

It would be cool if we could parametrize what module are present in assemblers!

[OminousZ]

I love it when devs assume the player is smart. It's absolutely the case.

[sakuro]

I am afraid that

Ingredient of: PARAM #NUMBER

part can be difficult to translate into some languages if the positions of PARAM and NUMBER cannot be changed. Will it be possible to reorder GUI elements via localization? For example, the most natural translation of this part into Japanese, which does not use prepositions, is

Code: Select all

[PARAM]の素材 #NUMBER

or

Code: Select all

[PARAM]のNUMBERつめの素材

(AのB is B of A and Nつめ is roughly equivalent to N-th).

[Viidi]

What about a parameterized factory? All this would be convenient to set up in a running factory.

Dynamic recipes, mmm...

[husnikadam]

Very cool feature. I have to also note, that limit of 10 parameters can be quite limiting for huge blueprints where the effort of filling out tens of parameters can still be worth than having more smaller blueprints. If you could find a way to make the parameters count unlimited, it would have been great!

[On_fire]

Just a thought, but it would REALLY be cool if we could somehow add a parameter to repeat the blueprint!
So it could ask you for the items in the blueprint, and one of the parameters could be how many repeats of the blueprint to make, potentially using a formula.

You could then make a blueprint that has 3 assemblers, and you setup 1 parameter to be for the assembler's recipe, and another to be how many assemblers you want with a formula to divide your number by 3 and round up... So if I know I need at least 10 machines for something, I can place down a blueprint, tell it I want 10, and it will give me 4 copies...

The repeat could be specified with custom signals for the direction, just like the parameter items.

[adam_bise]

I think this is great! I do hope Wube explores this sort of thing more. Anything that makes blueprints more dynamic!

[sb17joker]

This is great! I can't wait to use it!

[ryanalpasta]

You are a madman. I want this so bad now. I love tools, all the tools, who cares if I don't use it a ton? What's the harm in raising the skill cap of a game, especially pve? Even if I don't play this game All the time, this kind of thing is why I always come back and play again and why it's my favorite game

[Iceblade02]

Super cool idea tbh, can't wait to see what people do with it!