foliagecanine wrote:
eekee wrote:
(I needed Pygame for a balanced ternary computer emulator.)
Ternary? I'd like to see this...
![Very Happy :D](./images/smilies/icon_biggrin.gif)
Certainly! It's called by the world's most unpronounceable acronym, SBTCVM.
![Wink ;)](./images/smilies/icon_wink.gif)
I used the
Mark 2. I had some input, mostly "the CPU needs a stack" because the Mark 2 doesn't have one. (Such is the nature of learn-by-doing.
![Smile :)](./images/smilies/icon_smile.gif)
) I implemented stacks in software, but didn't get much of anything else done and ended up realising I needed to get better at low-level programming in general. There's a better SBTCVM now, which I haven't used because I drifted on to other things. It has a higher-level language which isn't Forth.
Gen2-9 The screenshots look impressive!
I quite enjoyed balanced ternary, but I wanted something a little more real and I came to realise it's not very practical to implement in hardware -- as far as I can understand. I figured out some logic gates but the best I can do is barely better than the RTL of the 1960s, and relies on particular transistors to even achieve that. (They need to be MOSFETs with a 7V turn-off voltage.) A Russian academic who was involved with the Setun balanced ternary computer explained its design was cost-effective in the era of expensive transistors. Perhaps it was, but it didn't make the best use of its core memory. It used a pair of cores to store one trit; holding 1 of 3 values, where a binary computer could store 2 bits in the same 2 cores; 1 of 4 values. This is strange because I would have thought the cores could be polarised either way to store 1 trit each, but I don't really know anything about how core memory works or how the Setun practically operated with diodes and coils in place of transistors. And I'm sure I'm waaaay off topic now.
![Smile :)](./images/smilies/icon_smile.gif)