"heavily influenced by https://biggieblog.com/optimizing-cost-in-opus-magnum/; esp. part about how to arrive at a lower bound to a solution; you need a bonder; debonder; rejecter; upgrader and the output themselves; and I think having a tracked arm in the same direction is *probably* necessary cause the piston is not on the same axis on the outputs." - Au, E.

don't fight molecule nature; placing a quicksilver divider is easier than flipping that output and try to wrangle the molecule into the other direction.

Influenced by https://biggieblog.com/optimizing-cost-in-opus-magnum/; and https://biggieblog.com/defeating-refined-bronze-and-losing-a-cow-bet/ where biggie demonstrated infinities in the alchemical board. 

omclone used for programming.

who said a debonder or a track is necessary if you have hands and can break the dowsing rod in two?

theoretically speaking; not considering any movement laws on atoms there could be a 26 on outputs; 1 on a piston arm; 2 each for a rejection, a bond and a projection glyph for a 35 area solution; but I highly doubt we'd see a 35 solution. I have no luck getting a 36 or 37 solution tho.