This is my first Forum post EVER and i want to try to make a contribiution for making our belts great again. As i`m not a nativ english speaker/writer, i would like to ask for forgiveness for spelling issues and stuff.
Belts in Factorio are in a rather difficult spot right now. They don´t have the throughput needed and suffer from UPS losses if you scale them up to get there.
So here is my idea. In my high quality Paint Picture it is to see what i would suggest as a possible solution for belts. Introducing a new belt which allows plate-like goods to be put on vertical instead of horizontal. This belt should explicitly only be usable for Iron and Copper plates as well as Plastic and Chips (mostlikely Green chips) maybe even for gears. Why this? Because for these high quantitiy products is where belts suffer the most. You need 16 or even more lanes of belts just for iron plates and the same goes for copper and in the end for green chips and the like. When you can increase the transportdensity of these on belts you can become competetive with bots. For the final products with"low"-quantity you can then choose to go with normal belt or bots.
Why do i propose this? Because most plate-like goods are also in high quantity need and are the most difficult to distribute with normal belts since you need a metric shitton of lanes for them ( Just saying 32 belt balancer for iron plates...) With a Vertical Placement Belt (VPB) you could have a belt variant with increased throughput for these goods but not for all of them. It is open to discussion if each segment should work like a seperate carrier so that no plates can move up till the end of the lane or if it should just work like a normal belt but just with more lanes to it.
The carrier option would have the benefit that each plate has a defined position on the carrier and no collision checks needed tobe made for all plates on the belt. But there would be the need for building a closed loop of these belt systems so that the carrier are always in motion or another solution for securing a carrier rotation on the belt. A possible option would be a terminator and starter like approach, where at the start of the belt (depending on flow direction) a starter element and at the end a terminator elemnt would appear. This would funtion like this: If a carrier is at the end of the VPE it enters the Terminatorelement and gets "transportet" back undergrund of the VPE to the startet elemnt and reappears there.
The up moving variant would just be a normal belt with more lanes to it so more checks on a single belt element. I don`t know what would be better performance wise
Also open for discussion is if a special type of inserter would be needed to load/unload the VPB but i personally think that is just unnecessary complexity.
Problems or at least difficulties with the VPB which needs to be adressed are:
For both variants
Placement of goods on the belt -> In which predictable manner will goods be placed on the belt? How will balancing work?
Belt Mixing -> Will it be possible to mix the lanes together? For example having a belt with two iron, two greenchips and a copper lane
For the VPE Carrier variant
How can the carrier return to the start of the point? -> Demanding a closed loop construction? Starter - Terminator element variant?
IF a Startet - Terminator is used as proposed in question 3, does the carrier need to be empty to be transported back?
IF a Startet - Terminator is used will the Starter and terminator be build automaticaly at beltplacement or should they be seperatly buildable elements?
I personally would favor the carrier variant since i think it can be a viable solution. I hope that i don`t have a error in my thoughts somewhere and if it so than please point me at it
PS: It`s a copy of my same ringing reddit post which i used as kind of benchmark to get a bit of courage to propose my idea here