[GTALUG] OT: From Nand to Tetris MOOC

[Lennart Sorensen]

On Tue, Mar 22, 2016 at 09:38:31AM -0400, Stewart C. Russell wrote:
> I liked the bit where they couldn't get the technical powerhouse of RCA
> to get the promised Selectron thermionic memory devices working, so they
> had to revert to Williams tubes. These had been running successfully in
> British computers since the mid-1940s. As with most early computing,
> seemingly the main difficulty was keeping the whole system cool while
> being able to discern signal from noise. That, and vermin control.
> 
> The IAS computer is definitely a better first computer to look at than
> the first stored programme digital computer, the Manchester SSEM
> (“Baby”). It had 7 instructions, could only subtract, and had one crude
> conditional. It takes real dedication to do anything useful with such a
> sparse instruction set. Still, people have implemented Baby emulators on
> pretty much everything — including the microcontroller scavenged from a
> low-energy lightbulb.

You only need one instruction:

https://en.wikipedia.org/wiki/One_instruction_set_computer

That will do it all.  It's awful to program though and certainly not
efficient.

-- 
Len Sorensen