On 27 Jun 2007 at 15:23, Billy Pettit wrote: Quite honestly, I didn't know what the time frame of the original question was--I did mention the Univac SS-80, however--and I did make reference to strange designs of the 50s. But "computer" isn't very specific. How about the hybrid computers of the 60s? How about the "mostly diodes and not a lot of transistor" machines like the PB250? Some parameters for the original question might be helpful. Cheers, Chuck

Billy wrote.... SortofKindaNotReally MicroData Reality machines were firmware implementations of Pick (Reality), not software implementations. As a result, you couldn't actually microprogram the thing as an end user (normally). There *WAS* an EEPROM board that would replace the firmware board, but I don't believe that was typically released to the field. I used to have two M1600's (the early ones with the wood panels). One of them did have the EEPROM microcode option but I do know that particular unit was sent out from microdata "under the table". I had let these two machines slip through my fingers many years ago, long before I got into collecting. BUT - on the bright side - I found out a year or so ago that those two machines actually did wind up in the hands of another collector who is known to another listmember! I now have an M1600 that is the later metal panel variety, waiting to be restored. Even on the software implementations (which the microdata's weren't), you could create new missionary instructions from native instructions by playing some games in the 5-pass assembly process. However, this wouldn't technically be considered microprogramming as the "native" instructions would be just whatever the underlying cpu was (68000, PDP11, RS6000, x86, etc.). So you could create new virtual instructions, but still not really microprogramming. Jay

It was thus said that the Great Billy Pettit once stated: In college [1], one of my bosses [2] worked on an analog computer. It was this huge hulking box, maybe 10' long, 6' high, and maybe 1 or 2' deep that you programmed by plugging wires into various modules. My boss used it to simulate a chaotic system [3] and play around with it in real time. Since I was never really good with electronics, I never did "program" the thing. Another computer I had access to (but sadly, never played with to any degree---sigh) was the Maspar. It was probably similar to the Connection Machine (single instruction, multiple data) but with only 4K CPUS, not the 64K the Connection Machine came with. And you could only program the thing in Fortran (which is probably explains why I or my friends didn't do anything with it). Ah, but the Connection Machine---now *that* was an interesting architecture. 65,536 1-bit CPUs that you would program in a varient of Lisp. -spc (But I know how to use a slide ruler ... ) [1] Florida Atlantic University, just across the street from IBM [4], where the IBM PC was first built. [2] I worked for a professor and his grad-student wife. Both had degrees in Psychology, but he worked in the Math Department, while she was still in the Psychology department, then later in a research department. [3] The forumas she simulated where: xn1 = ((A * yn) + B) * xn * (1.0 - xn); yn1 = ((C * xn) + D) * yn * (1.0 - yn); A, B, C and D ranged from -4 to +4 (but for each "run" were held constant) with the x and y range restricted to 0 .. +1 and plotted. It was much faster to vary A, B, C and D using the analog computer than it was on the SGI [5]. [4] IBM is no longer in that facility. http://maps.google.com/maps?f=q&hl=en&geocode=&q=Boca+Raton,+Fl… [5] Personal IRIS 4D-35. I was the one that wrote the software to display that equation on the SGI. I still miss that system, but not the hardware problems.