One of the tricks of engineering is to know what you need to be looking for. The entire Holocron project has been an exercise in belatedly realizing one thing after another that needed to be researched and calculated and worked out.
I was so sure I was on track to start slapping down components. Then while I was re-reading the pdf on the lithium polymer charge monitor chip I'll be using (a cute little SOT-23 package), I realized the programmed pattern of conditioning charge and monitor before delivering full charge, and the taper off to maintenance charge, were all going to play badly with the requirement of the rest of the circuit to suck power for the LEDs.
It took only a moment of googling to discover this was properly called a "Load Sharing" problem, and articles had been written and circuits proposed to deal with it. Meaning I need to add more components to the board. And as I discovered when I checked my Eagle libraries, meaning I needed to source the parts, get their footprints into Eagle...
(I also worked for a while trying to make the footprint work with the through-hole neopixels I'd found. Looked for a while at SMD versions, but couldn't figure out how to properly surface-mount on the back side of the PCB. Looked for a while for pre-wired strings and finally decided on a minimal breakout board that could be easily assembled into a string with ribbon cable. Once I've made my final decision on a supplier for that, too!)
And then there's times that empiricism is close enough. I've done so much trick-line rigging, I knew before I walked into the theater I could use span-sets (rather, some slings from my rock-climbing collection) as supports and quick-links as pulleys. I also knew I'd need a little weight to drag the line back down to the ground after we'd put it away the previous night. So with about twenty minutes to go I finally got around to stitching up a tiny sand-bag. Which I filled with....rice from the kitchen.
I don't call it impossible that there might be some strange interaction between rice, nylon, canvas, the steel of the old building, or whatever. Or some unusual factor like hungry bats or a lightning strike. If I was doing this rigging on a ten million dollar satellite you'd bet I'd check every wacky suspicion that occurred to me. But for this, for a two-week run of a lightly-attended show in a black box theater, none of the plausible failure modes are of high enough probability to worry about.
It's the holocron -- which will be available both as kit and fully assembled board to purchasers -- that I have to get it right. And more and more I think that with the first one I just plain got lucky.