#scary circuits
1 messages · Page 1 of 1 (latest)
hmm
i was thinking in mine it would present a signal parameter of the station that was successfully bitwise anded
Right, wildcarded pull, it's... within reach, I swear
but that might be difficult
Yeah idk how to get the non-symbol into the station name
well, it doesn't need to be in the name
Or hand-encode them all
well, actually maybe it does
well the encoding is easy enough actually
cause it's just a number from 0 to 2047
lol yeah, the encoding is fairly easy, it's getting that to translate to station name
I don't know how to even get a symbol from a number
I guess you could index the list
well you can use the signal param in the station name target, right?
yeah you can index the list
Are there collisions?
I feel like there's gonna be tons but I suppose if you send index on one side you can avoid them, hmm
Wow that's complicated, I'd like to see what it is
So like, RC materials, it has to be unique to RC&GC, RC&Plastic, RC&Wire, yeah?
jesus my brain was in a different place for a second, my friend and his gf trying to set me up with someone and that is entirely not the mindset i need to be in for this logic right now
RL comes first, this thread will be here lol
i am a fickle beast, i may seem like i know what i'm doing, but in another hour i'll be a caveman again and unable to bitwise OR a 0 and a 1
XD well, the problem's super complicated, the doing of it isn't so bad, it's the figuring out how the pieces fit
ok you are still in recipe land
but assuming you just mean that it's one of those materials, yes
so let me give an example since i have the combinator here...
That's a lot of mappings
it's just indexed/mapped, so it's easy
selector combinator OP
also 2048 / 5 (5 quality tiers) == 409.6
which is scary
and i don't want to think about it
you can't actually dynamically select an input in selector
2048 is too clean
honestly selector combinator is the worst. id rather use noSQL
i feel dirty saying that
ah i need all the inputs to be the actual number amounts
Oh maybe
is there as simple way to make that
i have touched lua like once
Maybe talk to Kaza about it
it's not 0 indexed, but i think given the context that's actually correct
uh i'm not sure how lets see
oh do i have to do a.... anything check and then pull out the indexed one
i don't have the selector combinator hooked up by a wire
that's why this was complicated
i told u, i'm an idiot
wait i do
yeah i have to match first
i had it hooked up to the wrong one
ok we good
Hell yeah, honestly I think this is probably something
I'm going to roll my solution but I'll do it in a way I can include this if you get it rolled nicely
I think it should be handled as is
*N is stronk
Wait is it and or xor?
Ohhh it's and, copper and thing==copper, right
There must be collisions?!?!??
collisions are fine because the game schedules
there are many stations in my setup that have the same name already for unloads, i just have to target the name and the game would determine which one to pick
Maybe
I spotted two collisions on my break
Fuck I forgot my coffee
Material #3 for product #15 collides with material #7 product 15. 3 tests true when 7 is sent
my coffee grinder...
And I think collisions are bidirectional
If you request 3 for 15 you'll get 15 for 3
i'm sorry tell me more specifically about those collisions
Maybe I misinterpreted how you're de-encoding them, but if you want to broadcast a request for material #3 in the giant list, for product number #15, you do 3&15, yea
Provider type 3 knows it needs to send something to destination 15
Except... that will activate for 7&15 too
And also, 15 will think it needs to provide something to symbol 7
Wait no, it doesn't work at all in that direction?
So I guess greater number must be requester?
xor is flat out as well... idk this is tough
>>> bitmask |= (1 << 5)
> bitmask
> 32
> >>> bitmask |= (1 << 15)
> >>> bitmask
> 32800
> >>> bitmask |= (1 << 7)
> >>> bitmask
> 32928
is_present = (bitmask & (1 << 5)) != 0
i'm working through some things, remember -- i'm an idiot
i am not limited to 32 items though, no
only by the max int
Yeah but if you're getting a symbol back...
You also have to encode all the combos up front, which you could do with a formula in a param, but still
Okay here's a simple version of the collision. Say you have only 10 symbols
If you want any combination of those two symbols, with and, it will also be a symbol
And there are multiples that fit
lets say you have 10 symbols
1&5, 1&7, 1&9, 1&11
and you want index 1 and 3
00000001010
Only 31 items
right, just add a lot of 0's to the left
but also the index's are usually much larger than 1 and 3
Yeah I don't think it's extensible enough 😦
Not THAT way
I think the idea is still something maybe
Encoding to symbol, maybe not, ands... maybe?
unfortunately i can't get good tests going yet either because the index's seem... whacky?
i'm getting slightly different quality results for the index's i'm picking and it's not consistent for changing + or - any constant number
O.O
oh, F, it's because when I do any minus math, i'm ruining the index cause numbers that are "0" disappear
my eyes are bleeding
but the only simple way i can imagine would not work with the current constant combinator index's, as it exceeds the max int when bitshifting
i would need a much reduced symbol count to make it work
i need to stop touching combinators
lmfao
i feel like i'm all cracked out
This is an incredibly tough problem all around. Peak science
Never stopped to think if we should
i don't think it's actually that difficult, i'm just not great at this stuff
I'm decent at this stuff and imo it's very hard
all of the things I am good at are pretty much everything that isn't details
Yeah the details are the devil here
yes details make my eyes bleed
The big picture isn't so bad to put together. We're missing a few tools that'd make these problems near trivial
The tools we have.... I think we have enough
i think probably the solution is somewhere inside doing something like a sha256 on some encoded value for truncation
wait no that would never work because that removes the utility of math from the encoding
I think this is the critical problem
yeah i wish i could just and symbols
and what it actually is, is ORing the symbols, then anding to check for existence
but there is no symbol1&symbol2 result
w/o a key anyhow
it would be nice if there was
i think if you worked with a much more reserved item set in the combinator
it would be doable by just bitshifting easily enough
idk, i give up. i don't know how i managed to get myself on this track
my brain hates me
great