Re: The joy of FORTRAN

On 3/8/25 3:04 AM, rbowman wrote:

On Sat, 8 Mar 2025 00:21:15 -0500, c186282 wrote:
 

On 3/7/25 2:23 PM, rbowman wrote:

On Fri, 7 Mar 2025 13:06:38 -0000 (UTC), Dan Cross wrote:
>

DEC in the 80s and 90s had a very forward-looking vision of
distributed computing; sadly they botched it on the business side.

>
Their entry into the PC business certainly was ill-conceived.

>
>
    Well ... the times ... EVERYBODY had to have a PC .....
>
    Unfortunately the market was already saturated.

 That was part of DEC's problem.  The Rainbows weren't completely
compatible with MSDOS software. Then they tried to lock everyone into
buying their diskettes with a proprietary format. Add to that the single-
sided siamesed drives and other quirks they were too far from what the
market expected by then.

   I do kinda remember the Rainbows ... and they WERE
   kinda weird.
   LONG back bought a Sanyo 550 ... an IBM *almost*
   compatible. Better base graphics, but not really
   CGA compatible. Bought a CARD that WAS CGA
   compat and, with a few jumper wires, got pretty
   much full IBM specs. Thing is, they weren't
   TRYING to defeat you, weren't TRYING to make sure
   you only used THEIR stuff. DEC *did*.
   In any case, "compatibles" were still relatively new,
   there were legal issues still in the courts, and as
   noted there were already SO many clones that the
   consumer market was saturated. DECs attempt to lock
   people into just THEIR stuff was a big no-go.
   IF the Rainbow's were like 10x plus times faster
   than typical clones for about the same price THEN
   they'd have had a real market. But, weren't ....
   The box I always wanted way back was one of the "Sage"
   line ... looked kinda like an IBM desktop but was actually
   680xx based. Several OS's offered. Apparently they
   had problems scaling-up production so they kinda faded
   away. LISA/MAC competition didn't help. Reportedly, early
   Amiga-OS was developed on them.
   https://en.wikipedia.org/wiki/Sage_Computer_Technology

    BTW, that page on PLCs was very interesting.

 Alive and kicking. A few years ago I interviewed a fairly young guy and I
was surprised when he had PLC programming on his resume. I'd been away
from industrial applications for over 20 years and thought they would have
faded away in the '90s.

   I *do* hate ladder-logic diagrams though ...  :-)

Relay logic was sort of fun. You were building state machines with relays,
Eagle Signal electro-mechanical timers, limit switches, and so forth. I
never did it but even more fascinating were some of the complex machines
that were all cams, levers, gears, and springs except for the drive
motor.
 https://www.hpmuseum.org/srw.htm

   I did some small stuff with actual relays ... but
   nothing super-complex. Control for smaller machines,
   maybe only eight switches/inputs. WAS kinda fun - and
   worked very well. The clickety-click of the relays
   always sounded "more real" than later electronics.
   My few later projects were microcontroller based, all
   software control, mostly 8051, but 'Rabbit' made some
   great boards (no longer human-usable after Rabbit-II
   alas because they kept shrinking the connections).
   Often was matching motor speeds or vehicle speeds
   to motor speeds - ag equipment. Found that the
   pain/strain of PID was oft unnecessary. Simple
   proportional combined with a little Fuzzy was
   often just as good. Best inspiration was to add
   a little look-up table of "what was working before",
   input rates -vs- PWM values. Quick to update, quick
   to use, minimal mem. So, when you engaged, it'd
   immediately use the last PWM value for the input
   pulse rate and then Proportion/Fuzzy for finer
   adjustment after. Made it snappy. Simple low-calx
   update/use of the table. Nematode-level IQ but
   it worked well.

With all its sophistication it wasn't protected against an attempt to
divide by zero. It would churn away until you pulled the plug and put it
out of its misery.

   Div0 has been a problem since forever. Fixes,
   if existing, have been kinda non-standard.
   Trying to GUESS if you are about to Div0 in
   the app tends to use up lots of code. Even
   now, not all processors have handy Div0
   exception handling ... I think even ARMs don't,
   or at least don't enable it by default. This
   is just *terrible* for 2025 ... I'd really
   not rather have my brain-surgery robot suddenly
   freeze up because of a stupid Div0 .