Is it a lost cause?

Really?!! I wonder if that's because there's one kind CPU in the world.
/BAH
Reply to
jmfbahciv
Loading thread data ...
Don't knock it. You could run your stuff on a PDP-8, turn the machine off, turn the machine on, and continue running your stuff.
/BAH
Reply to
jmfbahciv
It sounds like you have never done raw development.
/BAH
Reply to
jmfbahciv
band-aid
majority
to
certainly
high
worse.
them
I
TO
the
execute
times.
use
Not really.
Nope. Programmers wrote code on coding sheets and then submitted them to keypunchers (and later, tape prep types) who would type in the code.
You recall incorrectly :-). Read any *.MAC PDP-10 sources. We had writers who would document what the code did. Before the code was written we had product plans, project plans, design specs, and hardware prints, functional specs. There was an effort to get programmers to write extensive comments when they coded but it was a waste of time none of us had. Fixing code had to be done but fixing comments did not.
That was done in some shops. It depended on the site.
Well, I'd disagree with that. Reading C is no joy to me.
/BAH
Reply to
jmfbahciv
You set up the stack pointer by using IOWD on a PDP-10. Any register could be used but it was recommended not to use 0. Our "standard" for user programs (a.k.a CUSPs) used P=17.
A stack pointer could be switched with another stack pointer if you wanted to.
/BAH
Reply to
jmfbahciv
Then you have a different definition of commercial. I was in manufacturing.
/BAH
Reply to
jmfbahciv
That sounds like playing "whack the mole". I'd rather know what the instructions should be and then try to get something to generate the code.
Sure the bug should be reported but the bug won't be fixed and in the field before the first product ships.
That's a big IF.
So what did you do to make your code do the same thing under both?
I get nervous when I read that; it says that compiler coders are the most valuable people in the computing biz so they are a single point failure.
/BAH
Reply to
jmfbahciv
A KA-10 with 48K core did similar work in 1969.
/BAH
Reply to
jmfbahciv
Probably more than you have.
--
Outside of a dog, a book is a man's best friend. Inside of a dog it's  
too dark to read. 
 Click to see the full signature
Reply to
The Natural Philosopher
God, another troll.
--
Pete
Reply to
Peter Flass
Its becoming fairly clear to me that you have worked in entirely undisciplined environments where code has been hacked into place, with little design thought, or specification, no comments worth mentioning, arbitrary varaiable names and 'clever' but unintelligble programming style the norm from smart arsed 'permies' who dony have to be that professional....
--
Outside of a dog, a book is a man's best friend. Inside of a dog it's  
too dark to read. 
 Click to see the full signature
Reply to
The Natural Philosopher
In 1983 we ran a BBS that at the most had 34 logged on people on an 8Mhz 8086 with QNX as the OS. This was 16 via X.25, 16 via modems and one on the console and one on a networked (arcnet) pc.
Even running X.25 in software.
The CPU held up just fine, but the disk started to thrash a bit with these loads.
-- mrr
Reply to
Morten Reistad
It was a huge improvement over drum storage.
--
Steve O'Hara-Smith                          |   Directable Mirror Arrays 
C:>WIN                                      | A better way to focus the sun 
 Click to see the full signature
Reply to
Ahem A Rivet's Shot
Life isn't perfect, and sometimes a good brain can solve things where you cant see inside the black boxes.
Excatly. So you rewrite the code till it works
I've worked with highly imperfect 8 bit compilers back in the day. Often they didn't understand or made a hash of complex constructions. So I gave them simpler ones: If switch()..case: didn't work, if then elseif then elseif generally did...;-)
Nope. In general compilers for big systems like Linux and so on are pretty damned bug free and well developed.
Its compilers for piddly little embedded systems that tend to be less good.
rewrite it in different ways till the code worked in both environments
Well what about chip designers or chip microcoders?
The world is full of single point failures.
If CO2 levels in the atmosphere fell by a factor of 3, nearly all plant life on earth would struggle to survive.
formatting link

Its a damn good thing renewable energy is totally useless at reducing CO2.)
The Sun is a single point of failure too.
Compilers work, because by and large in they didn't all hell would break loose.
And then they WOULD get fixed.
--
Canada is all right really, though not for the whole weekend. 

"Saki"
Reply to
The Natural Philosopher
Your conclusion does not appear to follow from Barb's statement, indeed it appears to be in complete contradiction of it.
--
Steve O'Hara-Smith                          |   Directable Mirror Arrays 
C:>WIN                                      | A better way to focus the sun 
 Click to see the full signature
Reply to
Ahem A Rivet's Shot
wyse terminals too?
But exactly. The Sco UNIX PC was cheaper though.
I think that was when the per desktop cost of computing was at its lowest
--
New Socialism consists essentially in being seen to have your heart in  
the right place whilst your head is in the clouds and your hand is in  
 Click to see the full signature
Reply to
The Natural Philosopher
Nope. From 1983 to 1992 I was a full time software engineer writing embedded code for micros all over the place. 8080/z80/6809/8086 and friends..
I wrote a bit more after that, but was more system and network design after that.
--
"Socialist governments traditionally do make a financial mess. They  
always run out of other people's money. It's quite a characteristic of them" 
 Click to see the full signature
Reply to
The Natural Philosopher
That was the name I was trying to remember!
I thought I was later tole that its guts was ME29 hardware (soft as hell, apparently).
I was definitely told, when I started to use one on an ICL contract, that the 2903 used DFC hardware as the result of a "Christ, were short of cash! Gimme something we can make and sell quickly" panic by the new boss (Rob Wilmot?). So they shoved the DFC hardware into a new box, added microcode, a 5MB single platter drive (or EDS60 if you needed it - we did) and called the result a 2903.
Later, I remember the DFCs on a 2966 being a single, rather larger, cabinet. Was the 3 box solution an early version?
--
martin@   | Martin Gregorie 
gregorie. | Essex, UK 
 Click to see the full signature
Reply to
Martin Gregorie
There are actually quite a lot of them, they come in a small number of families but there is quite a gap between the lowest powered Atom designed for use in tablets and the biggest, fastest, multi-core-with-added -sauce-never-mind-the -power-consumption CPUs designed for gamers (or the similar but different ones for servers) - they all pick from the same grab bag of features with a common core but there is a lot of room for variations. Compiling to the common core should ensure code that works correctly across everything - but paying attention to the differences, especially the optional features, can make vast performance differences.
I think the main reason is that the regression test suites for compilers have become very deep and wide (and keep getting more so), and the resources are easily available to run them all on every build. By now this will force most bugs that escape them to be in new features, sane people give the compilers a while to stabilise before jumping on the latest language features for this reason.
--
Steve O'Hara-Smith                          |   Directable Mirror Arrays 
C:>WIN                                      | A better way to focus the sun 
 Click to see the full signature
Reply to
Ahem A Rivet's Shot
Same with 1900. The subroutine call used whichever accumulator you nominated to save a copy of the PC+1 and it was up the programmer whether he saved that (and any other accumulators) on entering the subroutine. Return was the exact reverse: you restored any accumulators you'd saved and the return instruction was told which accumulator contained the return address, i.e.
CALL 3 SUBNAME .... do stuff not using accumulator 3 .... EXIT 3 0
--
martin@   | Martin Gregorie 
gregorie. | Essex, UK 
 Click to see the full signature
Reply to
Martin Gregorie

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.