Page 1 of 3
Inserter drop point inconsistency
Posted: Tue May 02, 2017 1:18 pm
by Engimage
The well-known rule is that inserters do drop items on the far side of the belt if they do it from the side.
I'm not sure if this was always like that but currently (in 0.15) if inserters try to put items straight they always do put items on North/East side despite of their orientation
I will attach screenshot describing this.
Imo inserters should put items say on far side and to the right of their pointer so when you turn it 180 degrees it will start offloading to another side of the belt.
- Inserters direction.jpg (135.58 KiB) Viewed 7895 times
Basing on this intuitive logic I designed the marathon green circuit build
- Green circuits.jpg (166.96 KiB) Viewed 7895 times
But figured out it will produce only a single-sided belt
Re: Inserter drop point inconsistency
Posted: Tue May 02, 2017 1:35 pm
by Nemoricus
As a workaround, move the electric circuit removing inserter on the left down one spot and place a north-facing belt. This will let the electronic circuits go to the bottom lane. You'll have to move the medium power pole, of course.
Re: Inserter drop point inconsistency
Posted: Tue May 02, 2017 1:43 pm
by Engimage
Sure thing there are always workarounds. And I already did fix this in other design. However the inserter behaviour is what was bothering me.
After some research I came up with conclusion that when putting items straight inserters always put them on the right side of the belt so north/east is not actually true. The truth is that they still do use the same (right) side of the belt no matter of their orientation. So there is a predictable consistency in this but I'm not sure that it is desirable.
Re: Inserter drop point inconsistency
Posted: Wed May 03, 2017 10:24 am
by leoch
This mod lets you customise output location (and pull form the side). I consider it essential now.
https://mods.factorio.com/mods/GotLag/Side%20Inserters
Re: Inserter drop point inconsistency
Posted: Wed May 03, 2017 10:56 am
by Engimage
I know this mod. And I use Bob's one most of time. I am talking about vanilla here.
Re: Inserter drop point inconsistency
Posted: Wed May 03, 2017 11:36 am
by sparr
I, too, would love it if inserters were more locally consistent in the scenario you describe here. I've run into this problem when rotating blueprints before.
Re: Inserter drop point inconsistency
Posted: Wed May 03, 2017 1:16 pm
by leoch
Then I agree, but I'd like it even more if something like that mod were stock. (It could be just one key to toggle between near/far or left/right sides of the belt, and one key to rotate the pick-up point.)
Bob's version is just over-complicated (and over-capable).
Re: Inserter drop point inconsistency
Posted: Wed May 03, 2017 4:46 pm
by Koub
What's funny is that back in the old ages, the inserters did behave like you whish. It was changed for optimisation reasons.
I think it was with 0.12, but I might be mistaken.
Re: Inserter drop point inconsistency
Posted: Fri May 05, 2017 6:27 am
by BenSeidel
Koub wrote:What's funny is that back in the old ages, the inserters did behave like you whish. It was changed for optimisation reasons.
I think it was with 0.12, but I might be mistaken.
If that is true (the optimisation part, not the old behaviour part), with the new belt optimisations, is the inserter optimisation relevant anymore? Can it safely be removed to give us the old behaviour back?
Re: Inserter drop point inconsistency
Posted: Fri May 05, 2017 11:08 am
by leoch
BenSeidel wrote:If that is true (the optimisation part, not the old behaviour part), with the new belt optimisations, is the inserter optimisation relevant anymore? Can it safely be removed to give us the old behaviour back?
That's not the way code optimisation works.
In any case, removing a feature to improve performance (or something else) isn't optimisation really, it's simplification.
Maybe Koub meant the "optimisation" was from the UI perspective: simpler & easier to learn?
Re: Inserter drop point inconsistency
Posted: Fri May 05, 2017 11:50 am
by Koub
Wen I said "optimization", I meant :
Changes have been made to inserters and belts to improve the game performance (so that you could build bigger factories with same performance, or have better performance for a given factory).
One of the changes that came with these optimizations is the change in inserter/belt interaction.
Re: Inserter drop point inconsistency
Posted: Sat May 06, 2017 3:37 am
by BenSeidel
leoch wrote:That's not the way code optimisation works.
That is exactly the way code optimisations work. The original optimisation was based upon the original items-have-a-hitbox while on a belt, so there was a lot of work to be done to find a gap in the belt. As the belt mechanic has fundamentally changed, even more so with Rseding's new pointer-to-the-head system, it may not be an optimisation that will have any effect on game performance now as the inserter does not need to do an area search, but have a pointer directly into the belt datastructure?
Additionally, One optimisation may increase code speed, but in the presence of another optimisation, decrease code speed.
Re: Inserter drop point inconsistency
Posted: Thu May 11, 2017 10:13 am
by mOoEyThEcOw
BenSeidel wrote:leoch wrote:That's not the way code optimisation works.
Additionally, One optimisation may increase code speed, but in the presence of another optimisation, decrease code speed.
Exactly, without the code base, there is no way for us to know what optimizations do what.
Re: Inserter drop point inconsistency
Posted: Thu May 11, 2017 11:10 am
by BenSeidel
mOoEyThEcOw wrote:BenSeidel wrote:leoch wrote:That's not the way code optimisation works.
Additionally, One optimisation may increase code speed, but in the presence of another optimisation, decrease code speed.
Exactly, without the code base, there is no way for us to know what optimizations do what.
Hence the question directed to the developers:
BenSeidel wrote:with the new belt optimisations, is the inserter optimisation relevant anymore? Can it safely be removed to give us the old behaviour back?
Re: Inserter drop point inconsistency
Posted: Fri Jun 02, 2017 5:26 am
by gustaphe
+1 to this request. After 1337 (hey, look at that) hours in-game, something like this should be intuitive.
Re: Inserter drop point inconsistency
Posted: Fri Jun 02, 2017 9:32 am
by Lav
Placement being dependent on inserter/belt rotation is effectively a bug. It harms the gameplay, causing issues with an important game feature - ability to rotate blueprints.
Inserters should follow a specific set of rules. It doesn't matter what set of rules (though backwards compatibility is always nice), for as long as it's consistent and direction-agnostic.
Re: Inserter drop point inconsistency
Posted: Fri Jun 02, 2017 9:34 am
by steinio
Surprisingly 1 million buyers have no problem with the current functionality.
Re: Inserter drop point inconsistency
Posted: Fri Jun 02, 2017 10:46 am
by Lav
steinio wrote:Surprisingly 1 million buyers have no problem with the current functionality.
So following your logic, there's no need to fix bugs at all? Since, you know, a
million lemmings can't be wrong million buyers have already confirmed that they have no problem with them?
And how do you know they don't have a problem anyway? Have you asked all of them? Cause I can't remember answering.
Re: Inserter drop point inconsistency
Posted: Fri Jun 02, 2017 12:28 pm
by British_Petroleum
If this can't be fixed without making inserters more performance hungry, then I'd much prefer it to stay the way it is. It's quirky
Re: Inserter drop point inconsistency
Posted: Fri Jun 02, 2017 1:08 pm
by bobingabout
I don't really care. It wouldn't affect me either way if you fix it or not
This is... "Non-standard use of an inserter", I mean, they're configured for "far side of a belt", far side is the middle, so... non-standard. Besides, my inserters mod allows you to select what part of a tile to place items, even in this situation.