Conveyer Belt Master Race

[Mochan]

Something needs to be done about train/truck/drones.

Conveyer belts (and pipes) are simply the superior option to transport anything. The only reason to use the vehicle options is because it looks prettier, and looks awesome. Conveyer belts can get messy.

But Conveyer belts (and pipes) simply transfer things so much faster than freight or cargo, and require zero electricity/fuel to operate. This is a huge design flaw IMO.

In additions, I'm deeply disappointed with drones. I thought Drones would be a swarm of fast moving small aircraft carrying packages quickly to different buildings in their small flight range.

Instead I get a gigantic VTOL cargo craft with a tiny inventory that takes forever to lift off and land. It's ideal for long distances, but to be honest I think trains are better for that purpose. You won't need to worry about laying down tracks, but its output is so slow. I don't see much use for it, especially since by the time you unlock it you will already have established your Bauxite supply line, and probably your Uranium supply line as well.

[Vectorspace]

Counterpoint.

A single rail line can transport many times more resources than a single belt. It is quicker and easier to build one rail line compared to multiple belts.

You want to transport a new resource from A to B you have to create a new belt. Whereas if you already have a rail line near A and B you just have to add new stations connected to the existing line.

You can easily transport different resources on the same train, by having different wagons for different resources. To have different resources on one belt you either have to perfectly match consumption rates, or continuously sink excess

[kLuns]

The funny thing is that the throughput of trains and trucks is above the belt limit. So the throughput of any vehicle always gets limited by the belt.

[tulle040657]

My single freight car of rubber transports over 1100 rubber per minute. If needed I could add more cars to the consist

[Vectorspace]

Originally posted by kLuns:

The funny thing is that the throughput of trains and trucks is above the belt limit. So the throughput of any vehicle always gets limited by the belt.

But that's per wagon, and freight platforms have dual inputs/outputs. And multiple trains can share the same line

[tulle040657]

Yes a single freight station can output 1560 ppm

[Vectorspace]

Originally posted by tulle040657:

Yes a single freight station can output 1560 ppm

Slightly less, because the station stops transferring in/out during load/unload. I use both ports with mk5 elevators/belts to transfer to/from industrial storage containers which act as a buffer, then a single mk5 belt between that and my factory. That way I can guarantee 780/m even with the time the station is locked.

[Dampback]

Youre not doing it right then. Youre probably making to simple/short of a train run, Or youre not mass loading enough of something on there,

It will happen sometimes you need a small amount trucked or trained in. But having a belt (and all the items on it) will bog the game down, and if you go exclusively belts and avoid vehicles. you will get some lag.

Belts are definitely messier, but even so unless you put a lot of smart splitters in, its way mor efficient to have a train mass load from a to b and then do the splitters/filters from there.


----EDIT---- and the more wagons you have the better, plus the dual liquid and solid products on one line easier. PLUS rails transfer power, and I prefer all my stuff on one grid instead of a bunch of individual grids (comes with its own problems but easier to manage IMO)

[Wood_[OBC]]

long distance mk5 belts "lose" items, and therefore cannot sustain the 780 throughput. You "could" snap a splitter on every joint, and "weld" the belt segments together - but that creates other problems. The suggestion that vehicles are no faster than belts, is technically erroneous.
1 belt to station, then train, and finally 1 belt from station to usage = 780/min
however;
8 belt segments = 730/min
and it gets progressively worse, the more belt segments you use.
This is why vehicles are a more stable option for any resource transportation between factories imho

[Man's Best Friend]

In addition to other things already said, once you have some rails setup, adding another train is as easy as adding the end points, and the train itself. Much easier and faster than having to drag a whole new conveyor the entire distance.

Also tracks carry power to all connected stations, so that's another thing you don't have to separately bring to a new outpost.

[Grandaddypurple]

Originally posted by Mochan:

But Conveyer belts (and pipes) simply transfer things so much faster than freight or cargo,

Oh no I come back to this forum to read this...

As long as your roundtrip time is lower than [cargo capacity] / [input rate], everything is equal. If set up properly, your machines should finish eating up every items stored in the delivery truck / train station the SECOND the vehicle comes back to the delivery station
Trucks and trains do have the advantage of being able to replace several belts (one truck between 2 truck stations = one belt, one truck between 4 truck stations = 2 belts etc...). The only cons to trucks and trains are, as you mentioned, power usage and logistics, which are things the game was made to overcome (though to each their own)

[Mojo]

Originally posted by Mochan:

by the time you unlock it you will already have established your Bauxite supply line, and probably your Uranium supply line as well.

That is what I use my drone for now to bring in my Bauxite. I was able to get rid of the truck and 2 truck stations, and get faster deliver of my Bauxite using the drone. So for me the Drone is a win. :)

[ricky.brooks]

I'm not sure bugs and framerate issues as technical proof is a very good idea...

I think all bugs aside, if you were to optimize for max performance, you would technically be right.

You can also TECHNICALLY get the best performance in software from writing everything in machine code, even if it takes literally 1000x as many man hours and is completely unable to change due to a maintainability factor of 0.

These games are fundamentally about design, functional an aesthetic. You are claiming that the only benefit to using transport methods other than conveyor belts is the aesthetic design, but you are entirely neglecting the functional part.

You should be building systems that make it easy to expand, navigate, troubleshoot, and decipher, and developing design patterns that are easily repeatable, quick, hard to make mistakes, and easy to fix errors when they do happen. When you think "I need to connect these 5 manufacturers to these 4 lines of resources," you should feel that's no problem instead of being the biggest pain in the ass you can imagine.

Factory games VERY similar to software design, and you're suggesting that spaghetti code is best.

[Mojo]

Originally posted by ricky.brooks:

you're suggesting that spaghetti code is best.

I have seen some spaghetti code do better on things than a newer pretty code can do.

Just sayin, all depends on the application :D

(No ill intent meant, just pointing out we had a old program where I worked, after they "polished" it up and had "better code" it wasn't as functional as the old one.)

[cswiger]

Originally posted by ricky.brooks:

You can also TECHNICALLY get the best performance in software from writing everything in machine code, even if it takes literally 1000x as many man hours and is completely unable to change due to a maintainability factor of 0.

Optimizing compilers with peephole libraries usually generate better code than even fairly experienced human assembly programmers. Humans don't reason well about liveness analysis, CSE, and invariant code motion. They also have problems predicting and avoiding pipeline stalls which are more common with deeply pipelined CPUs.

Basically, the rare subset of people who can write really good assembly are better used creating the peephole optimization templates so that commonly used code idioms receive architecture-specific improvements by the compiler.

[ricky.brooks]

Originally posted by Mojo:

Originally posted by ricky.brooks:

you're suggesting that spaghetti code is best.

I have seen some spaghetti code do better on things than a newer pretty code can do.

Just sayin, all depends on the application :D

(No ill intent meant, just pointing out we had a old program where I worked, after they "polished" it up and had "better code" it wasn't as functional as the old one.)

That was my point. While it might "do better," or provide other short term benefits, there's a cost that's not being calculated here.