Pioneers of computing

[ben]

On 3/10/2019 9:11 PM, Will Cooke via cctalk wrote:

> 
> I have seen some claims that this was the first microprocessor -- although not a single chip
> https://en.wikipedia.org/wiki/Central_Air_Data_Computer
> 
> Will
> 
I would say it was JUST too early to count as valid microprocessor. I 
expect they all were 'hand picked' from the few chips that tested as 
working.

In some ways those designs seem better developed that the 'consumer'
computer products that came out. I would say the IBM 360 halted any real
progress since the 1960's with packing 4 8 bit BCD/text characters in a
32 bit word. I like 10/20 bits or 12/24 bits as a computer word length 
with byte addressing. IBM 360 32 bits 16 word reg file - 16 bit word.
PDP 11 8 word reg file 16 bit word. RISC 16 word reg file - 16 bit
word. RISC 256 word reg file 32 bit word.

Mostly the same format as reg to reg and load/store as the
be the model for most computer languages around 1970 ish. WOW a new
university computer from IBM ( 360 or clone) or a PDP 11 for the lab.

I was just reading somewhere , a  single user ? linux machine had 233 
threads going. What would that be like multi-user when it starts thrashing?
I think PDP8/e time sharing @ 110 baud got more real work done.

In hindsight, only after the fast 4K x 1 dynamic ram came out did
computing make it from the lab to the public with 16KB for OS
and 32KB+ for user programs. CP/M (8080) and FLEX (6800).

The lack of hefty card edge connectors like for the S100 bus
has me developing a 3 card cpu using 72 pin .156" pitch and
50 pin .156" pitch card edge connectors for a 12/24 bit
CPU. Emulated I/O planned is a TTY (terminal 1200 baud) and RK05
disc (PDP8) on a SDC card.  The time frame is 1975 ish
with the advent of 256x8 PROM's and 2901's and 74LS TTL with
250ns 4K DRAMS.

The bare machine is just burned today into a ALTERA DE1 FPGA development 
kit. Right now I am looking for few good books on a SIMPLE OS
and a SIMPLE programing language in the 1975 to 1980 time frame.
The catch is for now disk I/O is 12 bit words packed into 2 8 bit
bytes.

While it is TOO late to implement, a 8" floppy could be formatted
using GCR to give 8 512 (12 bit word) sectors per track using ballpark
calulations.
Ben.