Paul Bishop (paul.bishop.delete@this.bytemind.com) on July 31, 2022 4:08 pm wrote:
> Would be interesting to know your background and reasoning
> behind the research; self-interest, book, article, etc.

I am a retired programmer with nostalgia interests in hardware/software tradeoffs, like back in the 70's and 80's when minicomputer companies still designed their own cpus. I'm awaiting training to become a docent at the Computer History Museum. Until then, I practice being a docent online. I'm active on a private Facebook group called "Minimalist Computing", where fellow hobbyists chat about their hardware or software projects. I post there about the old computer designs when logic and memory were scarce.

Had fun learning about the IBM 5100 APL desktop computer of 1975. It was designed at a now-gone research lab that was a few miles from my home. Object code translation was my career specialty, so I tried to see how to convert its APL/360 binary to some efficiently executable form for a PALM-sized cpu. Instead of running multiple layers of interpreters. I couldn't even statically track the code flow, with all 360 branches depending on register values set elsewhere.

A group friend Doug Hale talked about about his boss (Bob Feretich) designing a very fast, minimal cpu when at IBM, the 5-bit controller. Described it as better than the famous 801 RISC, but rejected by IBM engineers for important (mainframe) work. That got me curious about its speed and design and uses. It has been interesting to see how much Bob remembers or can still dig up about his work nearly 40 years ago. I don't have that recall or records for my software work back then.