[kovarex] [16.0] Inserting on underground belt behaves different

[Aidiakapi]

In 0.16.0 the behavior of inserting on an underground belt has changed. I am not sure if this is intentional, but it was not mentioned in the changelog afaik.
The main downsides of this new approach are:

1. It no longer reliably compresses a belt (48 steel furnaces with yellow output inserter, where the last 6 inserters are inserting on red underground belt constantly fail to compress the belt).
2. It prevents undergrounds to be used for priority output, where it'll always prefer to get output from one machine, rather than another.

Old behavior: https://youtu.be/kVvhPCW1OQY
New behavior: https://youtu.be/08IH_2oeKms

Log for 0.16 attached, but don't really think that's useful.

[kovarex]

This mechanics was never intended and it was actually kind of a bug and using it feels like a weird trick.
There are other ways to get compressed belts (splitters for example) that should still work properly.

I'm aware that this possibly breaks some setups relying on the functionality, but we couldn't really change anything if we want everything to work the same all the time.

[Aidiakapi]

Understandable, although a well known trick among veteran players, it not very intuitive.

I've been doing a bit more testing, and it seems that out of the 3 ways to compress previously, only splitters remain.
Sideloading now also no longer compresses: viewtopic.php?f=7&t=54627

[Engimage]

Both changes make me really sad.
I can definitely say that a majority of balancers rely on sideloading and majority of production "columns" rely on the underground belts being able to compress the flow.

We can live with it but it is really sad. In this case maybe it is a good idea to review inserter behaviour so that it will naturally compress belts by inserting items in the "holes" while backing up the lane? And the same for sideloading?

[MrHick]

There is no way to compress the belt inline any more?
I can't find a solution for the smelters unless I go for chest+Loader(Mod).

Any advice is appreciated!

I really did not want to put any modded items in production, my base should be able to run without any mods.

[Bilka]

higher speed belt for the last few inserters. Actually compresses better than underground belts, even in 0.15

[PunkSkeleton]

About ⅔ length of the belt just move it to the side and start a new one. Tunnel the old one under inserters and at the end merge the 2 belts with a splitter. I have been always using that method for smelting and for blue belts it is even cheaper and I have never liked to rely on dirty tricks especially in a game that is in alpha (beta?) phase.

[Koub]

People on this topic will most probably be interested by this :
viewtopic.php?p=322677#p322677

[Avezo]

Would be better is inserters themselves were able to compress a non-underground belt by moving items forward/bacward a bit instead of making that way of compression impossible at all.

[Oozenthor]

This ->

Would be better is inserters themselves were able to compress a non-underground belt by moving items forward/bacward a bit instead of making that way of compression impossible at all.

I always thought it was super dumb that an inserter could not stuff an extra item in the belt, watching gaps move past an inserter with a part in it's hand just looks wrong.

Maybe give inserters priority loading if the belt in front of it is anything less than 100% loaded.

[SHADOW13]

how about, if there's at least "half" of the size needed for inserted to drop the item - it drops the item on the belt and anything that was there on the other half of the belt (or less) gets to hold up for split of a second ... (like joining traffic on the road)

wouldn't this better saturate belts with inserters?
the only problem then is what to do if you keep getting less than half of needed space ... maybe splitter+side loading?

or make inserters use similar mechanic as if it was side belt joining to main belt? that saturates

[MasterBuilder]

how about, if there's at least "half" of the size needed for inserted to drop the item - it drops the item on the belt and anything that was there on the other half of the belt (or less) gets to hold up for split of a second ... (like joining traffic on the road)

This is the kinda thing where it needs to either insert if there's any space, or not at all.
If you do it only when 1/2 item space is available, it's not going to be clear to users and will honestly look like a bug. I can see the reports coming in now "Insertions only compressing things some of the time".

[JareX]

This mechanics was never intended and it was actually kind of a bug and using it feels like a weird trick.
There are other ways to get compressed belts (splitters for example) that should still work properly.

I'm aware that this possibly breaks some setups relying on the functionality, but we couldn't really change anything if we want everything to work the same all the time.

This change really saddens me. I agree that having to use underground belts to compress is akward, but what I really don't understand is the reason behind having inserters work in a suboptimal way. There is no good reason for inserters to not fully compress a belt automatically. And if there is a reason for that I think the game should provide a reliable way for doing so, maybe adding a new entity like a "compressor belt" or a "sideloading belt". Still I think inserters should compress by their own, that is the most intuitive way for them to work.

[ignatio]

how about, if there's at least "half" of the size needed for inserted to drop the item - it drops the item on the belt and anything that was there on the other half of the belt (or less) gets to hold up for split of a second

That would make almost all compression problems go away. I think it'd be sad - coming up with ways to deal with bad compression is one of the many things that makes the game fun.

[Froilen]

Would be better is inserters themselves were able to compress a non-underground belt by moving items forward/bacward a bit instead of making that way of compression impossible at all.

I don't see anything wrong with that.
I see why underground belts don't compress any more.
maybe some creative ways to compress cool.
but new players would find confusing doing wierd stuff to compress the belts just because insertes "don't feel like working"

[loganb]

I hope we can keep the behavior of side-loading fully compressing belts. Otherwise, it is incredibly clunky to setup priority belts since they have to be split into two and recombined or else lose a few % of capacity.

Similarly, belt collimators (https://steamcommunity.com/app/427520/d ... 259884681/) become impractical for the same reason.

[LucidMoses]

Personally I would like a New type of belt that compressed belts in line. i.e. Fully compressing a belt on output until it ran out of in put and then leaving a gape big enough that a standard inserter (maybe only the fastest inserter) could fill in the gap and make it a fully compressed belt.

[Oozenthor]

An alternative could be to 'quantize' the item positions on a belt so no item gets placed 1/3 or .24567 distance from the last item. There is exactly ONE space to hold an item (or a gap to insert an item). The inserter could then hold an item until a free space presents itself on the belt and only then place it.

For example here is a comparison of an inserter placing 1 item for every 1.5 available positions on the belt:

Code: Select all

Time Now New
0     A   A 
0.5   A   A 
1         
1.5   B   
2     B   B
2.5       B
3     C   C 
3.5   C   C
4       
4.5   D  
5     D   D
5.5       D
6     E   E
6.5   E   E

You can see items 'B' & 'D' being held for 0.5 until the next open belt position passes. The end result is that throughput does not really change and only unity spaces are left open and available for further inserters to fully compress the belt. i.e. a second inserter can now place items between 'A' & 'B' and between 'C' & 'D' to fully compress the belt, as opposed to the current state of affairs where a second inserter can insert nothing.

[enigma888]

From my point of view underground belts and inserters work as they should work. Inserters should not be able to drop items in graps which are too small for an item. What worked before was a bug (in my opinion). (only we're used to abuse this exploid it's not a feature)

So what we are suppost to do is: to play factorio. It's about finding solutions to problems.
Here is one solution: Do it the "enigma way"!
Underground the main belt under the inserters. merge (with balancer) the belt from the last 6 furnaces. The Balancer compresses the belt!!

crompress-belt.gif (1.91 MiB) Viewed 18518 times

It's just a little addition to the 0.15 build and everything works fine.

[Zyrconia]

Sorry, I meant to post this sooner, but I've been indisposed the last couple of days.

I don't like the change in underground belts and do believe it is objectively a minor change for the worse. The old design, bug or not, had some advantages, like elegance, compactness and it was easy to understand. Additionally, the new change will break my whole mega base and it also widens the gap again between belts and bots.

The first point of elegance is very hard to quantify, but I do believe the old solution was more elegant. It also looked good and I started to use it everywhere, even when not needed, especially with mods. I even use it in my mining rows, together with a pole and a light in each hole.

Without this trick, we either sacrifice the elegance or the compactness of our builds. I wanted to explain this in words, in a lengthy paragraph, but enigma888 came to the rescue. He likes the change and illustrated a solution I wanted to suggest too, but used an animation instead of words. In this animation, we can see an alternative that first of all works correctly. Second of all, is IMHO less elegant, more noisy/denser and compared to adding a couple of underground downright clunky. Doing that is just busywork. Nobody want to worry with such stuff, we just want a saturated line. To be fair, the underground solution was busy work too. Ideally, inserters would be able to saturate lines on their own, but we need tricks for now. The smallest less busy work trick should win.

And it doesn't work with beacons. As far as I'm concerned, the pre-beacon phase of Factorio is probably the greatest and most fun mega-tutorial for building a mega-base in the whole history of gaming. But only with beacons will you have you final design for a mega-base.

With the old design, we needed exactly 7 smelters on one side to get a fully saturated blue belt. It relied on heavy underground use. It will take a very long time to come up with a new elegant design for this that is the smallest possible. After that I need to redo my entire smelting setup. I don't use a main bus and do not cross balance my smelting lines, because I use "better" modular designs, where everything is spreadsheeted and designed to perfectly accommodate full belts of input and consume them all. Even the smallest hole in my inputs will ruin my base.

Finally, underground belts are very easy to understand. Put two down. That's it! In 2 tiles you understood 100% of their complexity.

On the other hand spliters are more complex and have so many hidden properties. As a testament that they are harder to understand, you can take many random base, even from experienced players, and I'm sure you will find some redundant balancers, that actually do nothing. Or lopsided balancers, that do not fully equalize their not balanced inputs into s single output as the one who placed it expects to. Or if you need further proof on their complexity, do throughput analysis on various 8 by 8 balancers. See how easy it is...

If this change is final, I will surely accept is and continue playing Factorio, which is probably one if not my most favorite game. But I'll only update after I'm done with my current bases.