Improve Long Handed Inserters / Stack and Filtered Long Hand inserters / Increase stack, add filter to long handed inser

[MeduSalem]

My count exact, tick exact, slot exact contraptions heavily disagree with you.

Did you know that south facing inserters takes one tick longer than north facing inserters when inserting from box to belt? Well, I noticed because it spoiled the perfect lane alignment when creating a compressed belt by using inserters.

You know I agree with you on item exact, and also slot exact...

... But tick exact stuff sucks in this game (and any other game for that matter). Ever since XNight's Smart Furnace I despise such contraptions that have to be so damn deterministic that they choke on themselves if only a single tick misalignment occurs.

Yeah, I know that Inserters facing in different directions are acting at different speeds, but it has been barely measurable before people started to make crazy combinator contraptions that required tick perfect behaviour. Before that most contraptions still worked nicely and without any problems, including a Smart Furnace and various other things.

Now that a lot of circuit network designs depend on tick perfect behaviour to measure item amounts everyone seems to freak out because of that... but I don't really think it should go into that direction because in my opinion the circuit network was never truly designed to handle tick perfect contraptions in the first place... and neither do I think that a tick perfect solution is necessary to make item perfect contraptions - which is all that is really necessary to optimize factories and item flow.


You know what I compare that to? Classic NES speed running.

Why? Because those speed runners always complain about the game when it doesn't behave frame perfect like they expected it to... and they make their own life miserable because of their sense of perfection.

The funniest/best part about that is that it was never required by those games to play frame perfect to win. It was just required to react within a certain margin, within a certain tolerance limit.


So coming back to Factorio I think that people should forget about tick perfect control because it is way out of day-to-day usage scenarios that are imposed on players. As long as a perfect item count can be guaranteed then tick perfect contraptions aren't really necessary except if you want to make your own life more miserable. Perfect Item count is much easier to implement when done at the right spots like for example via Circuit Network adjustable Inserter Stack Size. Stuff like that's all that's really needed to gain back the control over the item flow we lost thanks to the Inserter Stacksize updates of 0.13.

I think a lot of people who are too deep into the Circuit Network stuff lost their perspective on the game as a factory game and that this factory game itself requires some tolerance and margin within things work... and that tolerance limit is basically reached with Item perfect setups, not tick perfect setups because you may need several ticks as a margin in which the entire circuit network is running because the circuit network needs to react faster to stay within the tolerance limit.


I like how both you and ssilk had that small discussion about quantum stuff in the "The sad State of circuit logic"-thread. It basically describes what I think... due to the way certain things within the game engine work you will always have this very small fraction of chaos even if this chaos is somewhat deterministic and even if it is only one single tick. So one absolutely needs a tolerance to cope with that chaos. What you expect is that the game itself should be reduced to 0 chaos so that the circuit network can be tick perfect, which is unrealistic to say at least without increasing the UPS to update faster than the CN updates so that you are able to catch states between states, increasing the resolution, etc.

Don't get me wrong, I'm for more control over the item flow, I just don't think that a tick perfect approach is the right approach... I don't even think that it is completely possible to make tick perfect contraptions reliable due to how the data is structured into chunks and how certain things work in a "cascade" fashion and whatnot. It's just not worth the trouble. Staying with item perfect on the other hand is a very likely thing.

One example I described here:
Only this combination is able to transport both lanes of the basic belt. If you replace basic belt with express belt, only one lane is needed.

Well, I admit it's a rare case. But I'm sure, there are more.

What the hell is that thing... and where would I need that? And why wouldn't splitting 1:1 with an express splitter be sufficient?

Well, when I remember back, I did it the same.
But now I'm really not for any kind of "automatic replacement". This gives the player a feeling of "this must be done to play the game 'right'". But it is sooooo expensive to use express belts. Or eplace basic belts with fast or - much uglier - express belts. IMHO in most cases not really needed. The problem is, that the game "looks like", that this is needed. But the truth is, that you make a faster progress, if you try to not use them and place parallel belts instead (where possible of course).

And that's why I came up with the belt-motor+modules that has an area-of-effect. No need to run around like a dumbass replacing belts anymore. Just plop the belt motor and the effect starts immediately... but the best is no one is forcing you to do it.

What is belt braiding?

Express Underground Belts braided with Fast Underground Belts and so on. That abomination:

Braids

fgeNwP1.jpg (105.53 KiB) Viewed 3209 times

[Yoyobuae]

Don't get me wrong, I'm for more control over the item flow, I just don't think that a tick perfect approach is the right approach... I don't even think that it is completely possible to make tick perfect contraptions reliable due to how the data is structured into chunks and how certain things work in a "cascade" fashion and whatnot. It's just not worth the trouble. Staying with item perfect on the other hand is a very likely thing.

You know a really nice aspect of tick perfect contraptions: synchronization. If one furnace works tick perfect, then the furnace besides that will also work tick perfect, with some delay due to belt propagation speed. But delays just mean a number of ticks.

And once 32 furnaces in a row work tick perfect processing a lane of a belt, then it's simply to mirror the setup to process the other lane.

Could even add another belt with 64 more furnaces, which would also work tick perfect. Everything perfectly synchronized.

Granted one has to account for the effects of chunk limits and such. But as long as those effects are predictable and accounted for, then it should just work.

[MeduSalem]

Good... the synchronization is a point in favor of tick perfect.

Back before 0.13 I just let my smart furnace run asynchronous... first come first serve style without a central controller... just a single wire from chest to inserter and it worked and I never had to care a bit.


I guess I have a different design philosophy and mentality about that... If I have to put a lot of maths into it to overcome game restrictions, which I also refer to as "gaming the system", I normally just say "screw it, let's do it the old fashioned but reliable way".

Maybe it makes me a loser because I never end up doing any cutting-edge stuff that would require bending the rules, but it also causes me less headaches and I will eventually cross the finish line too.

That said I used to be heads on too... But one gets older and the wear and tear did the rest.

[Roxor128]

Hm. It's useful to use slow belts in some (tricky) situations. An area effect would not allow that. You need to upgrade each belt separately. There are other ways to upgrade mass of belts.

I like the idea with the three basic inserters (burner, short
-arm, long-arm) and upgrade with modules. The hassle would not appear, if I can "mix" the currently needed inserter in my quick-bar together. Place belt and drag module over it (for example) and resulting item is changed.
But I fear it's still too much hassle.

Yes... I think being able to build a particular inserter configuration before placement could be just the thing to salvage my idea. Could be an extension of the blueprint concept. No, wait, even better, a precursor. Once you have upgradable machinery researched, you can create preset plans for building them with certain feature sets. You'd still need to hand-craft them to actually build it, though.

Maybe have a new tab in the build menu for presets. Design your own equivalent to a stack filter inserter, for instance, then build it whenever you need it. The game would build a short-arm inserter, a fast module, a stack module, and a filter module, then put the four together with one-second steps for adding the modules to the inserter.

Take the idea a step further and use these presets to make machines with modules already loaded and save yourself the step of going into the machine's menu and adding them when building a new line.

Express Underground Belts braided with Fast Underground Belts and so on. That abomination:

--image went here--

That makes my head hurt, and I've played 4D maze games.

[Deadly-Bagel]

Well, I admit it's a rare case. But I'm sure, there are more.

The one that comes to mind is a situation where Burner Inserters are preferred (usually either a reliable fuel-loading mechanism that can't lose power, or for loading ammo into turrets where there is no power) you NEED yellow belts because Burner Inserters are too slow to pick up an item from a fast or express belt. I've had cases where Burner Inserters try and fail to pick up from belts so many times that they actually run out of fuel and need to be started again manually.

As for belt motors, I hate the idea. How exactly would it work? Whack a motor down and it powers belts in a 10x10 radius? This is extremely inflexible in terms of managing belt speeds, such as belt braiding and a few other setups (the original splitter filter for example). Don't even try to suggest it would power a line of belts, many belts might only be a few tiles long and involve side loading. Belt lane balancers involve 3-4 short side-loaded belts that would all need powering. Then there's balancing, a motor powering 100 tiles of bus is identical to one powering 3 belt tiles used in a lane balancer. Not to mention it would be another convoluted mechanic for newbies to get to grips with, AND something else to clog your inventory, PLUS the added CPU cost of belts having to check they're powered (they're already the highest impact entity), ALSO belts now can't be used before you get power or in areas you don't have power, and at the end of it all what's the point? Just so the game is a bit more "realistic"? It's not like it would be a fun mechanic.

Anyway back on topic, yeah I could be on board with removing Filter Inserters. They're very rarely used in vanilla and it's not like they have a huge cost. Still, it's not like they have a huge cost so you can just carry around Fast Inserters and some iron (which you do anyway) and craft them when you need them. So it's so-so either way.

[thereaverofdarkness]

The idea of adding more types of long-handed inserters has been discussed before, and has been rejected likely due to the large number of inserters already in the game. Adding three new types of inserter just to give more long-handed options seems a bit overkill. But I have a better idea.


Advanced long-handed inserter
This inserter will cover all three bases at once.
Speed: slightly slower than fast inserter
Stack size: 1 higher than non-stack inserter
Filters: 1
Power consumption: 150kW
Power drain: 1.2kW
Hit points: 170
Color: pink
Recipe: 1x Stack filter inserter, 2x Iron gear wheel, 2x Iron plate
Technology: available with Stack inserters tech


One more inserter is still too much? I have an idea for reducing the current count without removing options:

Hybrid inserter
Color: honey brown
This is the basic inserter and burner inserter combined. It can accept either electricity or fuel. As it is faster than the old burner inserter, it will consume fuel even faster, so its power consumption is much higher when it is using fuel--20x perhaps. It also has no power drain when using fuel and prefers any fuel already inside it over electricity to operate, but will not put fuel into itself when it has electricity available.

The hybrid inserter covers the uses of both burner and regular inserters and simplifies things by allowing interchangeability. It also opens up a few new options for automation, such as inserters on a furnace line which will pick up coal from the belt for themselves in order to keep the furnaces running if power runs out. If your furnaces are supplied by a long series of ore and coal-filled belts, you could have the furnaces continue running for several minutes after a power loss, or perhaps you could power your mining drills with solar panels but without accumulators.

Another great thing the hybrid inserter brings would be code enabling hybrid machines. This could be later expanded into hybrid labs, assemblers, mining drills, vehicles, and more, adding more options in the early game as well as for bases that have difficulty providing power. Modders will undoubtedly also find interesting ways to provide hybrid equipment.

[JohnyDL]

You should probably try out Bob's inserters, there are long handed (even double long) inserters in there without new items, just by adding the ability to configure them (press E or shift+E while hovering on them I think) will let you chose the belt the side of that belt and more

[thereaverofdarkness]

You should probably try out Bob's inserters, there are long handed (even double long) inserters in there without new items, just by adding the ability to configure them (press E or shift+E while hovering on them I think) will let you chose the belt the side of that belt and more

Nice, I hadn't realized that customization was available when I was playing with bob's mods!

But it's difficult to use bob's mods just to get one specific thing they offer, since they add so many other layers of content. Even if you don't install the ores and you keep the material costs within vanilla types (very nice, bob!), it still alters the game sometimes more than a person might want.

[JohnyDL]

As far as I know all the modules in bubs are independent so you can use the inserters without everything else (I have before)

[MC_Cobalt_Ninja]

What about off-set inserters? Where it takes items from one space in one direction, and outputs it two in the opposite direction? Possible made from both a filter inserter and a long handed inserter. This could then also feature fast and filter capabilities, which would be pretty helpful.

[OdinYggd]

I already view the one type of long handed inserter that is offered as a crutch and build my entire factory without. It simply is not needed in the majority of use cases.

Adding more would just be silly.

[Koub]

The devs have stated long ago they didn't plan on adding new inserters. So you'll need mods for that, like :
https://mods.factorio.com/mods/L0771/Cursed-PI
https://mods.factorio.com/mods/Bobingabout/bobinserters

[bobingabout]

bobinserters (already mentioned) lets you configure an inserter after it has been placed. With 5 technologies to research (Long, Long2, More(rotation), More2 and Near) it allows you to configure the pick up and drop positions in an up to 7x7 (3x3 with no research, 5x5 with long, 7x7 with long2) grid, More research allows diagonals and more2 allows all possible positions in the grid, Near(Short for "Near side of belt" as opposed to the far side) allows you to configure where in a tile the inserter will drop in a 3x3 grid.
This is accessed via pointing to an inserter and pressing SHIFT+E (I think)

Boblogistics adds a button to the top of the screen to open a "pre place" menu. It's rather simple.
There are reasons why it is in boblogistics and no bobinserters, but I have had requests to move everything inserter GUI related to the bobinserters mod, and I can't say I disagree with that.
I also add faster inserters to my logistics mod, I'm not as keen to add those to the bobinserters mod because some people just want the GUI.


What inserters do I use where?
In the earlier stages of the game for example, I will actually have a furnace going with 2 belts down one side, and 1 belt down the other. This will have 2 belts on the input side, the near side containing coal and far side the ore to smelt, and the output on the other side. This requires the input item to be picked up using a long inserter. note: the near side inserter shovelling coal is often a burner inserter, ore input long inserter, output a standard inserter at this stage. The coal being on the inside lane is quite specific, because when I upgrade to electric later, the coal belt gets ripped out and the configuration is left with a belt on either side.
I configure inserters to be long inserters quite often. it's not often I will use other configurations, like 90 degree, or short in long out, or even near inserters. But I do find that now that I have the configuration options to do these things, I will quite often use them when the situation calls for it.


Trivia, Inserters have 2 speeds. A turn speed, and a reach speed. Long inserters have a superior reach speed, so in situations of say, a shuffle inserter (Configured so that you pick up from 2 tiles away, and place 1 tile away in the same direction. Basically swap the belt and inserters around so the belt runs directly alongside the factory with the inserter on the outside instead of inbetween), long inserters can perform quite well, and without looking up the numbers, I can't remember if it comes close, matches, or out performs a fast inserter. They are quite good at what they are designed for.

[ssilk]

Merged two other threads into this