And the verdict is?
I used the bog-standard default slicing settings for the Mars 2P in the Chitubox software. I was tempted to try changing things for finer detail, but decided to leave that until I know what I'm doing.

In the event the result was very good, so it was probably a wise decision.
The end 2 chairs in each row failed to print. I got only the support pyramids with no chair on top. But the 3rd chair from the end in each row printed perfectly, and all chairs in between.
* This suggests that the UV illumination is not even, and drops off suddenly at each end of the build area. This probably means that only the middle 80mm square area of the build area can be used for fine-detail models. This may be on my machine only of course (Mars 2P).
* edit: see next post.
I found the missing bits of half-printed chair lurking in the resin, partly stuck to the FEP film, but easily removed. This surely affects the idea of leaving the resin in the tank unfiltered for next time? They would clearly impede the first layer of the next print. So if doing that, it would be important to check that the previous print was fully formed with nothing missing.
I set the dimensions in Templot to match Exactoscale code 75 steel rail, and it worked out very well. The chairs can be threaded onto the rail without difficulty (much easier than threading Exactoscale moulded chairs), but then hold the rail vertical (as intended) with no wobble or play. As you can see in the photos, there is no infill below the key (this is an option in Templot), so that the rail is held entirely by the key against the rail web as on the prototype, and not by means of the rail foot:
The plug under the base of the chair gives it a lot more "heft" than other 4mm chairs (if you can use such a word for such a tiny thing) so that there is no distortion and the chair sits flat after threading (unlike Exactoscale chairs).
Dimensionally there has been about 1-1.5% shrinkage. I don't know where in the process this occurs, but I suspect in curing. The finished sizes are:
chair length: 4.79mm. Prototype 14.5" = 4.83mm (Exactoscale 5.08mm, noticeably larger than scale alongside these).
chair width: 2.64mm. Prototype 8" = 2.67mm (Exactoscale 2.79mm, noticeably larger than scale alongside these).
plug length: 4.12mm. Design 12.5" = 4.17mm.
plug width: 1.97mm. Design 6" = 2.00mm.
Some allowance for the shrinkage may need to be included in the files. On the other hand, the shrinkage on the plug is maybe welcome to ensure a clearance in the sockets. But we don't want too much because they will be setting the gauge. Some final tweaking of the dimensions is likely to be needed.
I'm attaching the .ctb file below for Charles and anyone who wants to try repeating my antics in an Elegoo. But you won't be able to make any actual use of the chairs until we have some timbers with matching sockets.
(This file is ready to use on the printer, it is already sliced and ready to go -- just save it onto a USB stick.)
p.s. I'm using the "ABS-like" Elegoo resin. Results with the standard resin may differ. Colour grey, results with other colours may differ.
cheers,
Martin.