The off-axis correction isn't really a very complicated calculation, but it does have limits. If you let the beam go off at 90 degrees to the vertical then it's travelling parallel to the surface of the resin - which is a problem (especially because the resin curves around with gravity and the beam mostly doesn't, so they'll actually get further apart). Much better to just limit it to a small deflection (I was thinking more like +/- 15 degrees) and increase the Z height to get a larger printable region. The mirrors will also have mechanical limits, of course, which will prevent really large angles.

Getting absolute dimensions right will be a matter of calibration and fiddling, as it always is. I'm not sure, but I suspect that if you set up the Peachy just about anywhere, have it build a 1" cube, and measure the edge lengths exactly then you'd be able to fully calibrate the system in software. Obviously a larger cube or running the same test multiple times will help, but at the end of the day you'll have to live with "good enough" accuracy (which is probably limited by how accurately you can measure the cube).