Though, as she doesn't yet have such a spell, she supposes she should probably...get to work preparing the spell.
And shoot off a message to the man who she has already lost all hope of not remembering as just 'oneiromancer guy' first and foremost - maybe Sebastien? de Gaulle? Is that it?
...'Sebastian deGaulle'. Close enough. He's in her contacts, for fuck's sake. She should remember this.
Oh, uh, one other relevant question, for your associate - how big would the crystal need to be? I figure I may as well busy my idle hands, and drafting is a fine thing to do it with.
--AM
...This is going to involve the hypothesized triangular grid, she thinks, if she wants to make sure every atom is pinned down just so. But she can definitely make that happen. Hm, how many powers of two is she going to need to iterate, anyway, once she's assembled the initial triangles from silver atoms...
...Oh, hm, and then there's the overlap at the merged edges to handle. She's not making triforces; if she's building bigger triangles from smaller atomic triangles, she needs to fill in the middle with triangles that are 1u less big. Otherwise the edges will get bumpy. Or could she...
...No, she really can't; that way lies Inescapable Collisions Hell. So figuring out how to interlink various sizes of triangle like some sort of awful, more spaghettiful cousin of Spacechem or Industrialcraft piping it is.
Let's see, one, two, three, four is two plus two plus two plus one, eight is four plus four plus four plus three, she needs to make seven? Fuck. Wait, would... Ha! Yes! Four plus three plus three plus three! And she can keep repeating that pattern until she's at the preferred size, then just stack vertically once she's done doing all the spinning she needs to do to make the hexagon!
Yeeessssss. Problem solved. She just needs to figure out the base rate, so it doesn't get jammed on the input level...
She needs - three times four is twelve two-triangles, four one-'triangles', four three-triangles. Right.
So she just needs to balance the Ag-splitter like so - tape the hexagonal board to the triangular one like that - hmm, now how is she going to trigger the looping...
It reminds her of the production puzzles, really. Limited space, after all.
Ooh, she could make a sensor on the hex board...
And if she makes the output also step that board forward one...
Well, she could probably make a counter in any number of ways, to be honest. But then - oh, that might be a problem. Hm. Because what she needs is a nested for loop, really. Not just a single incrementor. Unless she wants to precalculate exactly how many atoms of silver she'll need, which just seems excessive, even for her.
...Maybe if she did a fullness sensor? Yeah. She could do something with - reservoirs, and switching pipes...
...Could she, in fact, do that. Argh.
Well, maybe if she does some I/O sigil shenanigans.
...Oh duh, if you can hook the reservoirs up to release based on fullness and whether there's a signal being passed, you would have the proper behavior of accumulating (e.g.) 0111.1111.1111.1111 -> 1000.0000.0000.0000 -
...Motherfucker she hopes that the answer she gets back happens to be a power of two side-length in unit atoms.
...Wait, shit, and then there's the final assembly that she has to deal with these damn triangle problems in! Shit!
...Okay, she can make this work. She will make this work. She knows how to use a (...) to good effect. And if she wanted she could just... Oh, no, that wouldn't quite work either, but it's closer, at least. ...That means she's going to have a little bit of silver left over, huh. Hmm.
She pauses. Considers. Tests a couple hexagon-assembling shapes.
...So actually, how hard would it be to assemble a 2n-2 triangle with nothing left over? 3-3-3-2 into 6, but then...
...Yeah, 7-7-7-6 into 14, so she just needs to rebalance the input splitter - thank goodness she did the numbers in pencil - and there needs to be another assembler tower like so -
And the final hexagon has sides of length n, assembled from one triangle of side length n, four triangles of side length n-1, and one triangle of side length n-2. Or, at least, it worked with a 4-3-3-3-3-2. She hopes it scales. She figures it will scale.
...She's going to check with a 5-4-4-4-4-3.
...Thank goodness. It works. Now to just...
-- design an entire instruction set and selection of waldoes for the triangle grid, because she totally hasn't already made this hard enough for herself --
Well, that wasn't actually that hard all things considered, it's "translate one of six directions" and "rotate cw/ccw" and "pick up/drop", even if pick up and drop are sensitive to alignment - bonding and (hypothetical) debonding being emplaced on the sides of cells - or, no, actually she's gonna make that a mode switch, let the manipulators do it instead of cells -
Anyway. It's ready. Once she has measurements and has worked out a good point to do the unit conversion. And, well, acquired sufficient silver.
...She's going to be seeing assemblers in her sleep.
...Anyway, where was she?
She messages OS:
Incidentally, I have a hypothesis as to why h. sapiens lilim are, well, like they are.
Ritual invocations, encoded on the genetic level, expressed by whatever means Lilith-or-whoever chose. Possibly with a forcible propagation component in the gene-line - invoking the law of symmetry, maybe.
I may have to minor in org-chem and/or biomed, at this rate. With a soft sciences degree.
My CV is gonna be such a mess.