(Picture taken before final, printed, knobs were installed)
In general, the design process was good, the design itself good, the electronics and coding good enough. The modeling and casting succeeded, which is all that can be hoped for the first time out on a new technique. If I were to go around again on it, I'd be going into unexplored territory again on a couple places; on designing a 3d print with interlocking parts (well, more interlocking than the V150, or the 1/6 scale Claymore Mine I'm currently having printed), and on making a PCB. I've known for a while I've taken perf about as far as I can take it, and for the larger component counts, smaller footprints, and above all more reliability I need to move to PCBs.
I would like greater interactivity, but what I built already exceeded what was likely to come into play, literally; this was a demonstration piece for a table-top game, and probably was turned on and passed around for no more than a few minutes. And may not have been turned on since.
The Dialights are cute, but not really worth it. The visibility and visibility angle is lower than a standard indicator. Standard indicators domes would also mean I could shrink the footprint of the faceplate components a little and improve the overall fit.
I'd make a fresh 3d model, 100% printed, with a tight-fitting lid designed with pockets for neodymium magnets. That would ease assembly, and access for battery changes. There are several other smaller changes in the faceplate, including a larger relief in the shade hood, and a slightly larger footprint (there is space in the pouch for that).
The model would also be more detailed in the body, with better "battery compartment" detailing, some small raised ribs around the terminals, inset for the dataplate, and so forth.
The electronics would be on PCB of course, with the VFD soldered directly to the board instead of soldering each pin individually to ribbon cable (which was a real pain). I'm tempted to go with encoders instead of rotary switches for flexibility, but the main change here is switching to a double-deck so "OFF" actually works! Another change to the electronics would be a hand-rolled high voltage supply, so I could run the VFD brighter. Lastly, the power would be changed to AA instead of 9V, giving a much longer service life.
The speaker was underwhelming when packaged into the device. I'd have to experiment with this; the first idea is to use a true conduction speaker (with a proper speaker driver aka LM386 or similar) and make a speaker out of the printed shell.
But when all is said and done, the prop did what it was supposed to do, and there's no particular impetus to rebuild it.
Funny. It's something I learned while "debating" with Moon Landing Conspiracy buffs; when you try to "rebuild" a project using everything you learned since -- making use of new technologies and new techniques -- it becomes an equivalent job to designing it from scratch all over again. The main thing that stays in this case is the artistic concept, and all the time we spent developing that.
But that artistic concept is the most fun part of the project, so if I were to embark on a similar prop, I'd really want a chance to re-think that as well. Build one for a different period, or build a different device entirely. Because then you not only get to incorporate the lessons learned on fabrication methods, you also get to do some more creation.