How Big?

So how big is this project?   That is, will I ever finish?

You can't optimise without defining what it is you want to optimise (the objective function).   So firstly we need to define what "Big" means - is it:
  • area - that's important as I want one camera to see the whole state of the machine
  • component count - that's important as it's the main factor in cost and time
As most of the bulk components costs less than half a pence each and this project is going to be very time intensive to assemble, then I'm going to use the number of legs as my measure of Big.   Resistors and diodes have two legs, transistors have three legs.   The total number of legs is what has to be soldered together and is what dominates the area of the final board.    So when I've got the decision between two designs, I've picked the one that has the least number of legs.

I'm expecting 2 transistors, 12 signal diodes, one LED and six resistors per memory cell, 16 bit registers and about 8 registers.   That's 256T and about 6000 legs just for registers.

The ALU looks like 13T per bit slice at the moment, plus two for the output buffer, that's about another 256T (#legs not yet calculated).

Now if I could get the instruction decode, bus select and memory logic into less than 256T then that would be much less than 1024T overall, which would be a great achievement.

Will this fit on one plane?   I don't know.   I'd like it to be transportable, so no bigger than 0.6m x 1m (stuck to a decent bit of plywood with a carry handle).    So that's a max of 24mm x 24 mm per transistor (with associated LEDs, resistors, etc).   That doesn't seem impossible at this stage.   The ALU on breadboard is 50mm x 100mm and has 13T, so that's only two thirds of the area per transistor.  However, this doesn't have any space to describe what is happening, which is very important.