Cree Mount wire to board

I take it the schematics were all still hand drawn at that time? I guess CAD didn't take off until the 80s when PCs were commonly used in engineering. Duck soup on a computer. I think we had one of the expensive CAD systems using Apollo computers for entering the schematics. I took the net list over to the VAX and wrote a program to deal with the traveling salesman problem. One advantage of the 3M insulation displacement technology is that a run can be done with one wire and two wires can be used on a pin. So it was easy to cut out any pin or to add a new one onto any point on the net. The longer nets in the design did bring the VAX to it's knees. I thought they had preemptive multitasking, but apparently it still had issues that let one job hog the machine.

I wrote some text processing in a DCL "program". DCL was very powerful, but it has an inconvenience of having a very high priority. So it would hog the computer until it was done. I was told to run it overnight. Ok, but this was on a job where many users shared a limited number of com ports into the machine. So people would run port hog scripts to prevent their connection from timing out. Unknown to the powers that be or anyone else one of the many programmers wrote a kill-pigs job that would find such port hog programs and kill them. Unfortunately it also killed my work batch run. I had a hard time convincing an onsite DEC support guy that my program was not crashing because it was crap, but he finally looked into it and discovered the kill-pigs program. Management was not happy it cost them my time and effort, but were also impressed in the work of the programmer. My programs started working after than and we got the manuals out the door. This was the job where I wrote a ship's bell clock program too. Like I said, DCL was very powerful.

Most people don't like documentation and other "grunt" work, but I often don't mind it. You can still use your brain finding creative ways of dealing with it. Same with test. I think that can be a harder job than the design itself. Often test is left as an afterthought and requires a lot more creativity.

So it picks one pin from the net to name the net after? I guess that's at least slightly better than a random net name, or net1, net2, etc.

I don't recall any computer-drawn schematics in that day. The company did have mainframes and lots of Midi-computers, but they were not really up to real CAD.

I recall wham the mainframe folk (an IBM shop) tried to get Engineering to use the mainframes as well (versus the midi-computers). The whole thing died when I brought a box of pure binary cards up to be copied, and it choked the machine, triggering a loud argument between the IBM fan boys and the <other> fan boys; they didn't notice when I slipped out the side door.

The problem was that IBM usually encoded pure binary as EBCDIC coded hexadecimal, and simply could not handle straight binary. Unlike just about everybody else.

I do remember DCL, but didn't use it much.

In those days, it was originally all assembly code on the bare metal (no OS), slowly supplanted by Fortran (with an OS that served mostly as a loader).

C did not yet exist outside of Bell Labs and maybe some universities. But VAX/VMS eventually had a very good C compiler.

And a very good Ada83 compiler, which was ten times the memory footprint of the C compiler, a fair reflection of the relative complexity of the two languages, which may be one clue as to why C won in the market.

I write design memos to myself as a way of organizing my thinking. These memos often are later converted into principles-of-operation memos.

Joe Gwinn

Am 04.09.21 um 21:09 schrieb Joe Gwinn:

We had a PDP11/45 with RACAL-REDAC software. The guy who worked with this: I don't think that he was happy, but he got results.

Later there was RACAL Redboard on the PC. Some friends and me spent an Easter weekend to crack it. It had a laser-punctured floppy with one defective sector as copy protection. No problem for the Microsoft Code View in High Memory reserved for IO. You could trace it building a 128 byte buffer, writing it to floppy, reading it back and verifying the expected failure.

Probably the worst time investment I have ever made. The program was such a pile of ****. Oh, worse! Much worse!

Some time later I saw a 1*2" advertisement of an Oregon based company in EDN. We must have been one of the very first Orcad customers.

And now, when I use that !"§$% Altium Designer 21, I want DOS Orcad back. At least the user interface.

Cheers,

Gerhard

Joe Gwinn snipped-for-privacy@comcast.net wrote in news: snipped-for-privacy@4ax.com:

You spelled suicude wrong. (bridge diving).

That's a useful application.

I don't know Altiun, but the Orcad DOS keyboard interface was wonderfully constructed.

A colleague was using a Mental Graphics schematic/pcb tool, and adamantly asserted that you needed mice for schematics. I showed him how the keyboard interface worked, and he was amazed - and capitulated.

That was exactly my suggestion; solder to the LED/carrier/heatsink, then wire-wrap to the PCB.

So, what do you imagine, a surface-mount wrap post, rather than an inserted one, soldered in a PTH? Your idea, not mine; how well would that work?

I remember our company was considering a CAD system and Mentor was a new start up with a barely working product. The demo guy was using a touchpad and I'm not sure any of us had seen a touchpad before. I don't recall what that company bought if anything at that time. I do recall my manager was opposed to the whole idea. Why? Because to select parts they let you type while they narrowed the list (this was the days of TTL/CMOS/ECL SSI, MSI and LSI devices, so lots to choose from). He was either too embarrassed to admit he couldn't touch type (I seem to recall lots of people using two fingers then, so I don't get that) or he just didn't think he could manage it maybe? Dunno.

Otherwise the system was totally GUI, so no typing needed other than the parts search and I guess file names or whatever. It was not at all like a DOS system.

Oh yeah, this was the same manager who on my handing him my initial effort to use Avery labels to Zerox common components for hand drawn schematics saving adding all that detail by hand every time ('374, '244, etc) he said he would look into it and stuck it between some books on his desk. I never heard from him about it again. Maybe he just didn't want to change anything... ever. Must have worked for the railroad in his first job. They never change anything until they absolutely have to.

Joe Gwinn snipped-for-privacy@comcast.net wrote in news: snipped-for-privacy@4ax.com:

The term was and is MINI-computer, not "midi".

They "fell in" between PCs and mainframes.

They are "midrange computers", but they were and are properly referred to as mini-computers.

Not so fast there. There was such a thing as a midi-computer, falling above mini-computers and below full-size computers (read, mainframe). These are all marketing terms, and are of course wonderfully elastic.

These days, googling for "midi" anything will yield a flood of stuff about synthesizers for music and the like - unrelated.

Midicomputers were commonly used in process-control applications, and as such had extensive accommodations for I/O involving non-traditional peripherals. Like an oil refinery or a real-time simulator.

.

Joe Gwinn

Joe Gwinn snipped-for-privacy@comcast.net wrote in news: snipped-for-privacy@4ax.com:

There were PC and there were mainframes and the MINI came in between them and you term is marketing crap. Never heard it mentioned in CS studies. And I was around when Burroughs and IBM were doing mainframes and watched as PC ushered in and then IBM with their AS400 MINI computer. Never heard of your moniker in computing. That does not mean much though, because I did not get exposure to everything. (or I would surely already be dead).

"CS studies"? Therein lies the key.

I was using midicomputers like the SEL 32/55 in the mid 1970s, well before the academic discipline "Computer Science" was invented. My colleagues and I were all hardware engineers.

The CS of the day did not know of embedded realtime or process control and the like, and wouldn't know what to do with a SEL 32/55.

I later got a MSCS, in 1981. I was often the oldest student in the classroom, and had been making my living as an embedded-realtime assembly-language programmer for many years by then.

Joe Gwinn

CS has little to do with computers.

Yes, but some parts are quite useful. The most surprising subject was Computability Theory, as it served to warn me off of numerous easy-looking but impossible or at least intractable problems.

Joe Gwinn

I don't know where you went to school, but in 1979 those courses were already some years old at Melbourne University. The 3rd year project involved controlling a physical model railway with 30+segments, up to eight digital-controlled locos (RS232 control signalling on a 200KHz carrier superimposed on the track power) and 1000s of interrupts per second.

There was no compiler, no operating system, drivers or code libraries, just an assembler and an Interdata 7/16 (later called Perkin Elmer).

I'd like to see someone try doing that without any traditional "CS" subjects.

Clifford Heath

BSEE 1969 Johns Hopkins University. There were the first programming courses, of BASIC, but no computer engineering or science courses.

Yeah. Same story, different box. It later became Concurrent, a competitor to SEL. I have programmed similar boxes, but no longer recall the details.

What are you talking about? My colleagues were all doing exactly that back in the mid 1970s. I got my MSCS later, in reaction, because I noticed that roughly 2/3 of the degree syllabus would be useful, so I went for the degree, for the resume. This being after the fact.

Being bilingual was uncommon in those days, and very useful. It's probably less common today, as CS morphed into a branch of Applied Math, and few programmers know any assembler whatsoever.

Joe Gwinn

Don't know about that timeline; Gary Kildall got his PhD in computer science in 1972.

Well, you have a point there.

From Wikipedia: Kildall, Gary Arlen (May 1972). "Global expression optimization during compilation" (Ph.D. dissertation). Seattle, Washington, USA: University of Washington, Computer Science Group. Thesis No. 20506, Technical Report No. 72-06-02.

So the University of Washington did have a "Computer Science Group" (part of the Math Department?) in 1972.

But not yet true in the Baltimore area. Seattle would be too far for night school.

Joe Gwinn