[A13] A2B: conveyor belts & co. [v0.13.2]

[noone]

That's the design I have in mind - aren't they equivalent ?

I also build the individual "translations" of the new and old Selectors using the old and new systems. As you can see, the "old" selector is easy to implement with a 2x1 block in the new system, whereas replicating the "new" system with the old Selector is much more complicated (but certainly do-able - note the two selectors will have their selected items mirrored). This is definitely an argument in favor of the new system. Of course, the real question is : how often do one uses BOTH exists "2" in the new selector ? I think the answer is "not very often", which then suggest the "old" Selector is a more appropriate design. Actually, I might start a poll to see how often people use both "2" exists in the new selector.

I hope I didn't come off as demanding or otherwise disrespectful to your work and vision of what your mod should be.

Don't worry - that's the kind of discussions I like My driver really is to keep the number of belt elements to a minimum. I would hate to have redundancy, because it only complicate the maintenance of the mod (for which I have very little time), makes updates difficult, confuses players, and clutters the screen (which I hate).

So, I am really keen to get the few elements designed just right to fit most design needs - and the kind of discussions we are having is what I need to open my mind to designs I had not previously envisaged. I still haven't quite gotten my head around your 'soft selector' idea, but I will give it more thoughts.

Note that anyone willing to contribute to the mod directly is more than welcome to do so - that's why it is on Github, and that's why we have a separate account for it on the forum. That way, hopefully, the mod will outlast myself on here, and won't just die out If you want to design the soft-selector, I'd be happy to add it to the mod and see how it works out ...

Right now, my focus is on the underground system.

Edit:
Thanks FamousShoes for the feedback. You do make a very good point - now I remember why I thought it made so much sense to have a symmetric splitter in the first place
I won't revert things just now "because I don't have the time" (and will ponder the matter further). Part of the issue lies in the fact that I am yet to play with the mod in A10 and thoroughly test the system for myself.

In the meantime, I'd suggest using the combination ' selector+ merger = old selector '. But it looks like this "Selector" question will remain a tricky one for the foreseeable future ... 

[attachment deleted due to age]

[1000101]

You see your 2x4 is not the same logically as the 2x3 I use.  The difference is that the selector gets a second chance to select the item again which requires two inputs.

How my 2x3 works given two possible states:

Selector state unblocked (filter output path is clear):
The "input splitter" has 50:50 of paths to pick, in the event of the "wrong" path, the second splitter again has a 50:50 to pick the "right" path and send to the selector.  Occasionally it misses but the entire system is built as a giant loop so it will eventually filter in.

Selector state blocked (filter output path full):
The "input splitter" will always pick the unblocked path.  The second splitter has a chance to send again to the selector if it becomes cleared as an item passes through the configuration.

Why the proposed 2x4 is not the same:
It locks an item in the merge which may or may not be wanted while the splitter holds a wanted resource.  The 2x3 only locks a wanted resource in the splitter.

The system is designed around the principle that the destination of the selector has a small ouput chain, that is, it's dropping resources off right at work tables.  As a result, it doesn't want to drop a lot, but it wants to do it frequently.  That way the system can auto-feed the work tables "just enough" at a time to keep work flow while keeping pawns busy doing ANYTHING BUT hauling resources - after all, that's what conveyors are for.

Take a look at this thread for a working example of screenshots.

Note:  They system is 100% optimal at about 70% capacity due the designs spacing forcing items to "zig-zag" through it with the exceptions being when an output node is full and the selector is blocked.  However, there are multiple output nodes for the same resource (multiple tables want the same resource).  I have tried to optimize my production chain to minimize multiple same-drop nodes but that isn't entirely possible for some resources.  Also note, that each output node accepts only one resource (only steel, etc) with the exception of the kitchen freezer.

I'm not really sure you'll be able to satisfy all needs without breaking your minimalist design.

[1000101]

And @69, if that Rube Goldberg example isn't an edge case, I don't know what is. 

To be fair, this was the "simplest" non-blocking ring system I could achieve with the given parts. 

[noone]

You see your 2x4 is not the same logically as the 2x3 I use.  The difference is that the selector gets a second chance to select the item again which requires two inputs.

I still disagree - so let's settle this with an experiment !

I implemented my design in-game, and got 11 items out of 14 I sent down (80%). The exact sorting rate is in fact 75%, i.e. 3 out of 4 items get sorted out, and the last one goes through.

Experiment design: chunks fed one at time and not tight (i.e. the belt wasn't packed). Each exit is not jammed. And only chunks (1 item type) was sent.

You can improve the system further by adding additional "Merger - Splitter" rows just below the "Selector - straight belt" line. Each additional pair will reduce the fraction of mis-selected items by half. I.e.:

A 2x4 design gives you 75% selection rate
A 2x5 design gives you 88% selection rate
A 2x5 design gives you 93% selection rate
etc ...

Of course, you will never get to 100%, only exponentially close ...

Also to be noted: these fractions are upper-limits, when you feed ONLY the item you want to sort. Feeding other items might reduce the exact number depending on their frequency, arrangement, etc ...

When doing the same experiment under the same conditions with your design, I get the same output of 75% selection. Neither design can do anything regarding contamination by other items. Really, I fail to see any difference at all.

But by all means, feel free to prove me wrong - that's what Science is all about ! Do the experiment above under the same conditions, and I am fairly certain that you will too get 75% selection success with your design.


[attachment deleted due to age]

[1000101]

I'll have to do an A9 A2B vs A10 A2B but the core versioning won't factor into the experiment.

The problem I see with your design is that you can't achieve the zig-zag movement of items on the line which will force the path through the splitter at idea conveyor load which is about 70%.  This gives it at best 100% success and at worst 75%.  The other method can (as you pointed out) achieve a nearly mathematical 100% (assuming infinite splitters) and will always lock a (potentially unwanted) extra item.

I'll see about making a video to demonstrate it but this is really becoming a tangent and if I may skip tracks a little...

...Your mention of the github sources and being open to 3rd party additions.

I think this would be best solved on you deciding which will be the standard, then if you would be so kind, make the other as an additional mod to serve as a bit of a tutorial for others (such as myself) on how to make our own add parts.  I would certainly be willing to help in this capacity and I think this would be where my soft-select/overflow valve idea would fit.  I had a brief look at your source but I am yet to actually look into it.

[jefferyharrell]

Your preview screenshot confuses me a little. Does the mod do underground conveyor belts now? Cause if so that'd be amazing. I've wanted to use this mod in a game for a while now but I just haven't been able to figure out how to make it work for topological reasons. The teleporters are neat but they've struck me as a little too sci-fi-magic; if you have teleportation why do you need hauling and conveyor belts at all, I find myself thinking, which kinda takes me out of the game. But if conveyors can be run underground like power conduits are (and like heating ducts are in the RedistHeat mod) that'd be really terrific. I'd use the heck out of that, I think.

[Famous Shoes]

Part of the issue lies in the fact that I am yet to play with the mod in A10 and thoroughly test the system for myself.

I'd be willing to bet this will be an issue for many of us for a while, at least until the other great and essential/core mods have a chance to update to A10, e.g., EdB UI, PSI, and so on. I can muck about and do limited play testing for my own and others' mods, but playing seriously enough to get a feel for the UX so we can give useful feedback isn't really appealing as yet. In other words, if I were A2B, I'd hold off on any belt shaking decisions until other authors have caught up and we can all get down to the very serious business of proper play.

[noone]

Does the mod do underground conveyor belts now?

Not just yet, but soon (i.e. next weekend-ish). I need to test it Alpha10d, and run a few more sanity checks. These underground belts (only straight lines !) will be released along-side teleporters+receivers at first. If it all goes well, I will then remove the teleporters+receivers pair altogether, as both system are largely redundant.

And yes, they do behave like power conduits, i.e. can go under walls and tables and etc ...

[jefferyharrell]

That excites me tremendously. I'm really looking forward to your next release.

[s7jones]

Hi noone. Does this combine well with any other mods other than Mechanical Defence?

[Damienov]

Hi noone. Does this combine well with any other mods other than Mechanical Defence?

No clashes with all alpha 10 mod except for Tools for Haul mod. When a pawn equipping a cargo (from the tools for haul mod) using the conveyor belt, the cargo will be loaded into the belt along with its contents, unfortunately the cargo will get stuck at the conveyor belt exit.

Since both is pretty much a hauling mod, I guess you should choose one to use for now.

[s7jones]

Hi Damienov,

Thanks for your reply. That's good to know as well, but more specificially I was wondering if there are any mods that particularly benefit from the use of conveyor belts too, apart from Mechanical Defence 2 and the vanilla hoppers.

Thanks again.

[noone]

[...] I was wondering if there are any mods that particularly benefit from the use of conveyor belts too [...]

We had a nice partnership with Rikiki's "Deep Driller" mod a few Alpha's back. But that mod was discontinued. At the present, the A2B belts can be used to feed the teleporter device designed by Haplo. It is still a transport mod, but as the teleporter/receivers in A2B will soon be removed, using the two together will become more meaningful.

Other than that, I am sadly not aware of any mod that could significantly benefit from the belts. I.e. I don't know of any 'machine and devices mod' out there. Ultimately, because the A2B belts can reload hoppers and any storage spot, chances are that A2B will work right away with any processing machine similar to the nutrient paste dispenser (e.g. only requires power and food to work).

Hope that helps. Did you have any specific idea/wish in mind ?

No clashes with all alpha 10 mod except for Tools for Haul mod. When a pawn equipping a cargo (from the tools for haul mod) using the conveyor belt, the cargo will be loaded into the belt along with its contents, unfortunately the cargo will get stuck at the conveyor belt exit.

Thanks for reporting that ! This i the first time A2B clashes with another mod, and is good to know. At this time, I've read here and there that the "Tools for Haul" mod has memory leaks, so, I would not recommend using it. But in principle, I don't see why this mod should not be compatible with A2B: I'll take a look eventually, and see if we can reconcile them.

[A2Bcorp]

We at A2B are pleased to announce the pre-release of v0.10.1 of the mod, introducing underground belts !

These much awaited and much requested belt elements will allow you to send items underground for a while (straight lines only !). The new 'Undertaker' and 'Undercover' pairs are intended as replacement for the Teleporter and Receiver pairs. Both systems are present in v0.10.1, but our teleportation system will be retired in the next release.

v0.10.1 is tagged as a pre-release, because the underground transport time is at present incorrect (i.e. fixed, no matter how long the underground stretch is). We are working hard on fixing this, but this process is taking more time than anticipated. Nonetheless, the mod is fully usable and should not give rise to any issue (restarting a game is probably required). As usual please report all issues on this list.

Final note to existing mod users: a pair of 'Undertaker' MUST be connected by a series of 'Undercover' element at every location between them. These can be build almost anywhere, and under existing buildings, walls, etc ...  But unlike teleporters, you won't be able to ship items under mountains anymore !

[1000101]

New additional selector components as a mod add-on available now, see the OP!

Gives you a total of 12 selectors for all your Rube Goldberg configurations!

Original thread [locked]: A2B - Selector Options