Re: Version 0.16.16
Posted: Wed Jan 10, 2018 11:58 pm
Hah this is spot on!Koub wrote:Honestly, to all those who want the unintuitive black magic powered splitters :
https://xkcd.com/1172/
Page 3 of 6
Re: Version 0.16.16
Posted: Thu Jan 11, 2018 12:08 am
I understood how splitters worked, saw usefulness in it, and appreciate the change to a simpler functioning.
If for no reason other than if your guy could make a splitter that complex, he should be able to just make a sorter
If for no reason other than if your guy could make a splitter that complex, he should be able to just make a sorter
Re: Version 0.16.16
Posted: Thu Jan 11, 2018 12:14 am
5thHorseman wrote:I understood how splitters worked, saw usefulness in it, and appreciate the change to a simpler functioning.
If for no reason other than if your guy could make a splitter that complex, he should be able to just make a sorter
Re: Version 0.16.16
Posted: Thu Jan 11, 2018 12:24 am
But that wouldn't be as cool as http://i.imgur.com/8sBmW0S.gifv5thHorseman wrote:I understood how splitters worked, saw usefulness in it, and appreciate the change to a simpler functioning.
If for no reason other than if your guy could make a splitter that complex, he should be able to just make a sorter
Re: Version 0.16.16
Posted: Thu Jan 11, 2018 12:26 am
Like thisFactorioBot wrote:Changes
- Changed splitters so they work more intuitively. The left and right lane splitting is now completely independent. The decision whether item goes to left or right output is now independent of the itemtype.
Re: Version 0.16.16
Posted: Thu Jan 11, 2018 12:55 am
And another feature bites the dust.
I don't see why these changes are bundled together, there's nothing prohibiting the per-item nature, and still having lanes separated. Separating the lanes avoids the issue people complain about, the per-item behavior is what facilitates the sorters and priority splitters.
The performance argument cannot really be that compelling, compressing the representation (say 30 bits item id, 1 bit per lane state), vectorize lookup, and you can compare 4 items in constant time @ the size of 2 pointers, which can double as discriminated union to a hashset to get guaranteed constant time at any item count. More than 4 items, even with mixed belts, is highly unlikely. I doubt you'll even need 30 bits for item id.
You could have improved belt sorting and priority splitting whilst fixing the issue, but instead opt to just remove it altogether, a trend in the 0.16 pipeline. Just because many people don't understand why something works, these features allow for very varied gameplay for those who enjoy that. And it's not like this is a case of "this weird thing that nobody knew about", it's been featured in your own FFF.
I don't see why these changes are bundled together, there's nothing prohibiting the per-item nature, and still having lanes separated. Separating the lanes avoids the issue people complain about, the per-item behavior is what facilitates the sorters and priority splitters.
The performance argument cannot really be that compelling, compressing the representation (say 30 bits item id, 1 bit per lane state), vectorize lookup, and you can compare 4 items in constant time @ the size of 2 pointers, which can double as discriminated union to a hashset to get guaranteed constant time at any item count. More than 4 items, even with mixed belts, is highly unlikely. I doubt you'll even need 30 bits for item id.
You could have improved belt sorting and priority splitting whilst fixing the issue, but instead opt to just remove it altogether, a trend in the 0.16 pipeline. Just because many people don't understand why something works, these features allow for very varied gameplay for those who enjoy that. And it's not like this is a case of "this weird thing that nobody knew about", it's been featured in your own FFF.
Re: Version 0.16.16
Posted: Thu Jan 11, 2018 4:46 am
I am very sad to see the splitter change. This broke most of my mixed belt setups. The outputs from the splitters are now very unbalanced (for mixed belts) because they do not retain the per-item splitting. Honestly, this change just pushed me from having fun with belts to just going to bots. I understand the independent lane splitting. I am fine with that. But, the loss of per-item split is a deal breaker. I love using belts and focused mainly on using belts. I am still going to try and find new solutions, but the ones I have been finding are very unsatisfying. However, I am very happy about fixing compression issues with splitters. That is great news! I just wish we could have the per-item split along with the compression fix.
Re: Version 0.16.16
Posted: Thu Jan 11, 2018 8:45 am
I haven't tested the splitter changes yet, but what I'm reading here, has me concerned. Not because of the change, but the expectations of some people here: This game is the dev's game. It is an alpha build - Things are subjet to change. It might break your stuff, but there is always another way to get to the goal. Yes, we can make suggestions and voice our opinions about feature XYZ, but we have no right to force anyone (especially the devs) to change it like we want.
If the splitter changes bother you this much, either play the version before this one or use mods. The devs have shown time and time again, that their judgement of what is healthy for the game and what is not, is really good.
Maybe the splitter changes will stay, maybe they will be reverted or something else will happen. Who knows? I believe, that the devs will find a reasonable way. Trust them.
I for myself welcome the splitter changes. From the looks of it, it makes perfect sense to not have it be item-dependent.
If the splitter changes bother you this much, either play the version before this one or use mods. The devs have shown time and time again, that their judgement of what is healthy for the game and what is not, is really good.
Maybe the splitter changes will stay, maybe they will be reverted or something else will happen. Who knows? I believe, that the devs will find a reasonable way. Trust them.
I for myself welcome the splitter changes. From the looks of it, it makes perfect sense to not have it be item-dependent.
Re: Version 0.16.16
Posted: Thu Jan 11, 2018 11:14 am
Caine wrote:http://www.nooooooooooooooo.comFactorioBot wrote:Changed splitters so they work more intuitively. The left and right lane splitting is now completely independent. The decision whether item goes to left or right output is now independent of the item type.
Another coolfeaturebuglostfixed
Re: Version 0.16.16
Posted: Thu Jan 11, 2018 1:59 pm
I fully agree.Schorty wrote:I haven't tested the splitter changes yet, but what I'm reading here, has me concerned. Not because of the change, but the expectations of some people here: This game is the dev's game. It is an alpha build - Things are subjet to change. It might break your stuff, but there is always another way to get to the goal. Yes, we can make suggestions and voice our opinions about feature XYZ, but we have no right to force anyone (especially the devs) to change it like we want.
If the splitter changes bother you this much, either play the version before this one or use mods. The devs have shown time and time again, that their judgement of what is healthy for the game and what is not, is really good.
Maybe the splitter changes will stay, maybe they will be reverted or something else will happen. Who knows? I believe, that the devs will find a reasonable way. Trust them.
I for myself welcome the splitter changes. From the looks of it, it makes perfect sense to not have it be item-dependent.
Re: Version 0.16.16
Posted: Thu Jan 11, 2018 7:55 pm
Well, at least under the right circumstances they are sorting better than ever but I have to admit that it is a lot less useful. Or even an annoyance if you are inputting repeating patterns and want to end up with the same ratio on both sides.
but I have to admit that it is a lot less useful. Or even an annoyance if you are inputting repeating patterns and want to end up with the same ratio on both sides.
- splitter sorting 2.gif (393.45 KiB) Viewed 7558 times
Re: Version 0.16.16
Posted: Thu Jan 11, 2018 8:49 pm
It is written a little bit too emotional for my taste, but to content I agree.deef0000dragon1 wrote: Dont pull things from the game like this. Find out WHY people are using it, and make it obsolete instead. With barrels, people use them because the current fluid system is asinine and broken. Dont break their capacity and all the current working builds that people have put SOO much time into. Make the fluid system functional instead, THEN balance. The same is true for bots. Why are people spamming bots? Fix the system, then nerf. Again for the splitters, why were people using the splitters like this? solve that problem of why, then remove what you consider a bug. Belt compression: Why are people using undergrounds to compress? Because things should be compressing naturally maybe? find out why, solve that problem, then remove the "bug".
You claimed over the holidays that you felt you were in the final stage of ballancing/polishing the game. you're not. You still have many things that need to be completely re written, and aren't just a numbers ballance.
It is most times better first to understand WHY something is done in a special way and then to do anything. Also I think it is more important to give alternatives then to prohibit (or remove) a common way.
But in this special case it doesn't hurt me much. It takes a lot of time to understand how splitters work ... and now it is exactly so like I have taught the first time I used it.
For so far .... good job devs.
Re: Version 0.16.16
Posted: Thu Jan 11, 2018 11:32 pm
I'd really like a device like this. Most of all it feels natural, as each input lane is mapped uniquely to an adjacent output lane, so it's simple to see/expect you'd get full throughput.Merssedes wrote:About underground belt lane splitters...
IMHO, to make lane splitting the best way is add item of 1x1 tile size that will have 2 input belts and 2 output belts next way:I1 and I3 -- lanes of 1st input (2 lanes)Code: Select all
O1 O2 I1>^ ^<I2 I3>v v<I4 O3 O4
I2 and I4 -- lanes of 2nd input
O1 and O2 -- lanes of 1st output
O3 and O4 -- lanes of 2nd output
(It's also kind of like sideloading two belts headed in opposite directions, if you could offset belts by half a tile)
Re: Version 0.16.16
Posted: Fri Jan 12, 2018 8:33 am
Finally splitters are working properly
Re: Version 0.16.16
Posted: Fri Jan 12, 2018 8:55 am
I have drawn similar thing in suggestions topicteleksterling wrote:I'd really like a device like this. Most of all it feels natural, as each input lane is mapped uniquely to an adjacent output lane, so it's simple to see/expect you'd get full throughput.Merssedes wrote:About underground belt lane splitters...
IMHO, to make lane splitting the best way is add item of 1x1 tile size that will have 2 input belts and 2 output belts next way:I1 and I3 -- lanes of 1st input (2 lanes)Code: Select all
O1 O2 I1>^ ^<I2 I3>v v<I4 O3 O4
I2 and I4 -- lanes of 2nd input
O1 and O2 -- lanes of 1st output
O3 and O4 -- lanes of 2nd output
(It's also kind of like sideloading two belts headed in opposite directions, if you could offset belts by half a tile)
viewtopic.php?f=6&t=56248&p=333579#p333579
Re: Version 0.16.16
Posted: Fri Jan 12, 2018 8:55 am
Actually, if anything, that seems more broken than it was before. I would expect both sides to end up with both copper and iron.Mimos wrote:Well, at least under the right circumstances they are sorting better than ever
Re: Version 0.16.16
Posted: Fri Jan 12, 2018 9:09 am
I think the current new behaviour is more logical, and more easily understood than the "splitters have indefinite inner memory that will remember where went the last item os one sort to ensure the next one that will come, be it in 100 years, will go on the other lane".
Now it's : whatever arrives in a splitter goes one right, one left, one right, one left. That's so much more logical for a mechanical splitter.
Now it's : whatever arrives in a splitter goes one right, one left, one right, one left. That's so much more logical for a mechanical splitter.
Re: Version 0.16.16
Posted: Fri Jan 12, 2018 9:24 am
I fully agree with you here. Have in mind that splitter is a very early game item and it should not have any complex logic it it. Remember that Factorio is a game for creating complex things from simple blocks and black magic definitely defies thatKoub wrote:I think the current new behaviour is more logical, and more easily understood than the "splitters have indefinite inner memory that will remember where went the last item os one sort to ensure the next one that will come, be it in 100 years, will go on the other lane".
Now it's : whatever arrives in a splitter goes one right, one left, one right, one left. That's so much more logical for a mechanical splitter.
Re: Version 0.16.16
Posted: Fri Jan 12, 2018 9:52 am
I still think your formula is fundamentally flawed. viewtopic.php?p=332448#p332448FactorioBot wrote:Tweaked the balancing of the PvP production score.
Re: Version 0.16.16
Posted: Fri Jan 12, 2018 10:13 am
True. What we need now are "smart"-SplittersKoub wrote:Now it's : whatever arrives in a splitter goes one right, one left, one right, one left. That's so much more logical for a mechanical splitter.