Page 3 of 4

Re: Version 0.16.25

Posted: Tue Feb 20, 2018 12:32 pm
by Tekky
PacifyerGrey wrote:
Tekky wrote:Also, it is inconsistent that filter inserters can be connected to the circuit network but the newly introduced filter splitters cannot. Therefore, it would be nice if this suggestion were implemented before 0.16 is declared stable.
This is a suggestion, namely a feature request.

These have nothing to do with game stability. Even less with defined release like 0.16. Devs might decide to change those or not. And definitely they won't oblige themselves to change it before 0.16 stable.
I agree that new features should generally only be expected to be implemented in major releases (such as 0.17) and not minor releases (such as 0.16.26).

However, due to the fact that an inconsistency was introduced in an experimental version, I stated that "it would be nice" if this inconsistency were resolved before 0.16 went stable. I did not state that it "should" be resolved before 0.16 being declared stable. In fact, I would find it completely understandable it this feature were delayed until 0.17, especially since the feature would likely not be trivial to implement.

In contrast to that, I believe that the issue of walls sometimes not connecting to diagonal cliffs is an issue serious enough that it should be fixed before 0.16 is declared stable, because it concerns an already existing feature and could be classified as a bug, because it behaves in an unexpected and unintuitive way.

Re: Version 0.16.25

Posted: Tue Feb 20, 2018 2:23 pm
by Engimage
Sadly provided compression solution is still buggy.
There is a bug posted by AntiElitz viewtopic.php?f=182&t=57998 which exposes inserter misbehaviour (while not game breaking)
And another more serious bug with sideloading shown here https://gfycat.com/FinishedEqualLeafwing

Re: Version 0.16.25

Posted: Tue Feb 20, 2018 2:26 pm
by dee-
Drury wrote:Hold on, if this does what I think it does, does it mean you can supercompress belts by putting items one over another?
It think does not do what you think it does.

Re: Version 0.16.25

Posted: Tue Feb 20, 2018 4:16 pm
by Zyrconia
Thank you! Great change!

But I will miss having to put undergrounds on output points. Because of visuals and symmetry, I used to put them everywhere, even when not needed. And it even helped UPS in 0.15.

Re: Version 0.16.25

Posted: Wed Feb 21, 2018 12:21 pm
by Deadly-Bagel
dee- wrote:
Drury wrote:Hold on, if this does what I think it does, does it mean you can supercompress belts by putting items one over another?
It think does not do what you think it does.
It doesn't, I mean to say that the Inserters just drop items into any available gap even if it's not big enough, and while the items do "supercompress" it's only until the belt starts moving.

These issues with sideloading seem to be that belts have a wider tolerance of compression than Inserters or graphics - Inserters work, and you can see gaps, but sideloading doesn't see the smaller gaps nor does the belt when it holds items to make room for a newly inserted one.

Re: Version 0.16.25

Posted: Wed Feb 21, 2018 12:27 pm
by xng
I'm super grateful for the compression patch, it's not perfect yet but releasing this means we're getting there! <3 <3

Re: Version 0.16.25

Posted: Wed Feb 21, 2018 7:01 pm
by PaszaVonPomiot
Can you show the difference in belt compression now and before? How much does it impacts the game?

Re: Version 0.16.25

Posted: Wed Feb 21, 2018 7:35 pm
by kaliszowy
Servers are down?:( I cannot download latest version.

Re: Version 0.16.25

Posted: Thu Feb 22, 2018 2:18 am
by Oktokolo
Hmm, compression by moving the items and holes of the belt backwards?! That looks ugly. I'll just keep using splitters. Compression is easy to achieve without the new cheat anyway.

But i really apreciate the fix for the search box in the mod download GUI. It was so annoying to search for anything before the fix. Now it just works.

Re: Version 0.16.25

Posted: Thu Feb 22, 2018 12:32 pm
by Nick-Nack
I'm really sad that compression was introduced again. In my opinion it goes against what Factorio stands for: figuring out best builds on your own. It's not that full compression is actually needed for playing the game or that you can't achieve it with some thought. It isn't even the case that creating a fully compressed belt with splitters takes way more space than an uncompressed one. So please remove autocompression again! I don't see how it could fit into what your vision for the game seems to be.

Re: Version 0.16.25

Posted: Thu Feb 22, 2018 12:57 pm
by Deadly-Bagel
PaszaVonPomiot wrote:Can you show the difference in belt compression now and before? How much does it impacts the game?
Would be interesting to see - sometimes I've had belts that are 1/2 item gap every 15 items or so, hardly a big effect, other circumstances I'd see maybe on average 4 items every 5 slots. I might try testing when I get home if I'm awake enough.

Though I think that if items aren't compressed, the belt mechanics treat each gap as a break - meaning instead of one long continuous flow, it's calculating lots of smaller ones. I could be wrong but this could also have a minor UPS improvement. Of course, it's also possible the fixes consume more UPS than that so it could also drop...
Nick-Nack wrote:I'm really sad that compression was introduced again. In my opinion it goes against what Factorio stands for: figuring out best builds on your own.
There are intuitive ("good") problems and unintuitive ("bad") problems. A "good" problem is something like the trickiness of handling three fluid outputs from a Refinery array - how do you tie each together in an expandable manner? You look at the tools you've got and work out a solution, awesome.

Belt compression was a "bad" problem, it's something you expect to work but it then doesn't. You track it back and find your "inserters" don't actually insert. The worst part is you can't actually see this problem until after you've got a full setup working with exactly a full belt, and you are also consuming all of that full belt. By this time it's probably much too late to change anything.

Re: Version 0.16.25

Posted: Thu Feb 22, 2018 1:33 pm
by PEBKAC
Any idea on when 0.16 becomes the main branch instead of Experimental?

Re: Version 0.16.25

Posted: Thu Feb 22, 2018 2:50 pm
by Aeternus
Belt.zip is a go! I repeat belt.zip is a go! :D
Thanks for the update.

Re: Version 0.16.25

Posted: Thu Feb 22, 2018 4:44 pm
by torham
PaszaVonPomiot wrote:Can you show the difference in belt compression now and before? How much does it impacts the game?
I have build my last factory before the "Uncompress" patch, and I haven't touched it until now. The production after this patch went up by roughly 10 - 15%. Its about compounding gaps. Gaps in iron mean less resources down the line, which means less intermediates and less final products. On top of that factories which would otherwise be unable to produce ( like green circutis) due to the belts being almost, but not quite full, can now work at full speed. Example:

I had a stack of 16 green circuit factories, and before this patch 3- 4 factories were always stalling due to inability to place on the nearly full belt. After the patch, the same stack goes full speed, and I was able to place another 4 factories on top, because I can better control the input of raw resources with configurable splitters.

Of course I understand that all of these instances can be engineered around, and input/output methods redesigned, but this is the difference between the two versions of the factorio.

Re: Version 0.16.25

Posted: Thu Feb 22, 2018 4:55 pm
by Lubricus
Deadly-Bagel wrote:
PaszaVonPomiot wrote:Can you show the difference in belt compression now and before? How much does it impacts the game?
Would be interesting to see - sometimes I've had belts that are 1/2 item gap every 15 items or so, hardly a big effect, other circumstances I'd see maybe on average 4 items every 5 slots. I might try testing when I get home if I'm awake enough.

Though I think that if items aren't compressed, the belt mechanics treat each gap as a break - meaning instead of one long continuous flow, it's calculating lots of smaller ones. I could be wrong but this could also have a minor UPS improvement. Of course, it's also possible the fixes consume more UPS than that so it could also drop...
Nick-Nack wrote:I'm really sad that compression was introduced again. In my opinion it goes against what Factorio stands for: figuring out best builds on your own.
There are intuitive ("good") problems and unintuitive ("bad") problems. A "good" problem is something like the trickiness of handling three fluid outputs from a Refinery array - how do you tie each together in an expandable manner? You look at the tools you've got and work out a solution, awesome.

Belt compression was a "bad" problem, it's something you expect to work but it then doesn't. You track it back and find your "inserters" don't actually insert. The worst part is you can't actually see this problem until after you've got a full setup working with exactly a full belt, and you are also consuming all of that full belt. By this time it's probably much too late to change anything.
I disagree I think the belt-compression was an intuitive problem. You could watch how the things on the belts and the inserters move, and figure out that the belt couldn't compress due to caps that is to small for one item to fit. The solution with undergrounds didn't on the other side feel intuitive. And the autocompress inserters is quite complex and non intuitive (how does the inserters stop the belt or even temporarily reverse it?)
I am also a player that don't like to plan to carefully and think it's more engaging to find out and fix problems under the construction. So I like problems that pops up that I couldn't foresee and mess up the design so I always have something to improve, that the main drive to play the game.

Pipes on the other hand have some bad unintuitive problems like that the update order influence the throughput and that pipe loops can hinder flow. That should be fixed.

Re: Version 0.16.25

Posted: Thu Feb 22, 2018 5:37 pm
by Deadly-Bagel
How would you observe Inserters failing to insert into small gaps until you have anything other than big gaps? I find your reasoning unsound. And it's not like the workaround was particularly exciting, just output to 2 belts and merge them together at the end.

Also, items don't "move backwards" on the belt, they overstack which then evens out when the belt moves. Eg in a 3-item slot, there are 2.5 items due to one item only being half-inside that space. Inserter drops an item into the 0.5 space - you now have 3.5 items in that 3 item space. The front items start moving, the item half "under" the inserted item stays where it is until there is room to move forward. It makes as much sense as the belts themselves do.

Re: Version 0.16.25

Posted: Thu Feb 22, 2018 10:22 pm
by Deadly-Bagel
Deadly-Bagel wrote:
PaszaVonPomiot wrote:Can you show the difference in belt compression now and before? How much does it impacts the game?
Would be interesting to see - sometimes I've had belts that are 1/2 item gap every 15 items or so, hardly a big effect, other circumstances I'd see maybe on average 4 items every 5 slots. I might try testing when I get home if I'm awake enough.
FURNACES
I threw together a setup of three furnace columns, one for each belt type, smelting Iron Ore. All Inserters were connected waiting for a signal to acivate, which would work for 20 seconds (to fill the output belt) then begin measuring for 60 seconds. Here are the differences between the two versions:

0.16.24
Yellow Belt: 739
Red Belt: 1,293
Blue Belt: 2,173

0.16.25
Yellow Belt: 800 (+8.3%)
Red Belt: 1,597 (+23.5%)
Blue Belt: 2,400 (+10.4%)

It's most noticable on red belts, which for some reason is 3 items off perfect (I guess I caught it on an off tick?) but yellow and blue belts notice a significant improvement too. Note that different setups will see different results on 0.16.24 - for example if you're using an Efficiency + Speed Beacon moduled setup then your Inserters are going to be in different places dropping items at different times, so they're going to end up in different places on the belts with different gaps. This gives a pretty good idea though.

I've uploaded the save as well, just switch on the Constant Combinator.

Re: Version 0.16.25

Posted: Fri Feb 23, 2018 1:56 am
by Oozenthor
Changes

Inserters and belt sideloading can now squash item on belt even when the gap isn't big enough. The squashed gap is extended to normal size once the front of the belt starts to move again. This means, that inserter rows and side loading can produce fully compressed belts without the usage of splitters.
Thank You.
Belt Furnaces

Re: Version 0.16.25

Posted: Fri Feb 23, 2018 2:21 am
by ignatio
Lubricus wrote:I disagree I think the belt-compression was an intuitive problem. You could watch how the things on the belts and the inserters move, and figure out that the belt couldn't compress due to caps that is to small for one item to fit. The solution with undergrounds didn't on the other side feel intuitive. And the autocompress inserters is quite complex and non intuitive (how does the inserters stop the belt or even temporarily reverse it?)
+1. Figuring out how to achieve good belt compression was fun. Now the inserters will use some strange magic to cram in items where they shouldn't fit. Less fun.

Is this an attempt to make belts more competitive against robots? If so it's sad, and also far from enough. The way to balance (read nerf) robots is imho to make their loading and unloading into hotspots. (Or don't - I don't care that much that robots are imbalanced. Don't use them for high-throughput anyway because that's just boring.)

I agree that this is a very un-Factorio-esque change, in that it takes away an efficiency problem for players to solve. Anyone is welcome to play with badly compressed belts - that works too, but if you spend effort on making them more efficient then you should be rewarded for that.

It's not often I disagree with a game mechanic change, and I have much respect for the devs and their feeling for high playability, but please reconsider this.

Re: Version 0.16.25

Posted: Fri Feb 23, 2018 4:09 am
by sathill
Great update. Seriously guys just focus on bulding big efficient base that produces this 1rpm (and others science). There its so much to do that thinking how to compress belt was lame (2 lines and splitter wow very hard)