[TR] Cheers!

Dave dave1massey at cs.com
Wed Apr 19 17:05:11 MDT 2017


The first company I worked for where I did embedded programming had recently upgraded from such a system.  I was told that an assembly took 18 hours.  We had a deep culture of patching where you would put in a jump statement to an unused part of memory, put your new code there and jump back to the end of the skipped code.

We were programming the old RCA 1802.  What a neat chip.

 

 

Dave Massey


 

 

-----Original Message-----
From: Randall <TR3driver at ca.rr.com>
To: triumphs <triumphs at autox.team.net>
Sent: Wed, Apr 19, 2017 1:17 pm
Subject: Re: [TR] Cheers!

> In the late 70's to 81 I was working with a mini computer that used core
> memory.  This was a 19 inch rack type device and the core planes filled
> several levels.  Then they came out with a really big computer that had a
> hard drive.  It was a good 14 inches in diameter and 5 or 6 inches tall
> and the annual service was to refill it with nitrogen (or some other
clean,
> dry gas).  It had a whopping 1 megabyte!

Me too, except the hard drive was a whole 2Mb, with half of that in a
removable cartridge.  It took up two 12" high spaces in a 19" rack, one for
the drive itself and another for the power supply.  That was only on our
development system though, the systems we sold had only 9 track mag tapes
(and sometimes not even that).

I tried once doing a build on an older machine that had only an ASR33
teletype for mass storage.  That was an interesting waste of time!  The
machine had only 8k of 16 bit words, so first you had to load the editor
through the paper tape reader (using the front panel to control the loading
process), then you could edit the program a few lines at a time, reading as
much into memory as would fit, making changes, then punching new tape.
Lather, rinse and repeat until it was right.

When you were done with that, load the first pass of the assembler, and feed
it your source.  It would produce an intermediate paper tape (which included
the source as well).  Then load the 2nd pass of the assembler and feed it
the intermediate tape.  It would spit out a relocatable binary file (if
nothing went wrong of course).  Then if you wanted a listing, you had to
load the 3rd pass, and feed it the output of the 1st pass again.

All of this took place at the whopping speed of 10 bytes per second.  Yes,
bytes.

Now load the 1st pass of the loader and feed it each binary file in the
correct order.  (I forget now why it made a difference, but the order was
important.)  It would spit out a symbol table.  Then load the 2nd pass, feed
it the symbol table tape and then all the binary tapes again, in the correct
order.  If you were lucky, it would spit out a paper tape with your
executable program on it.  (I wasn't so lucky and after having spent all day
on it, gave up.)  Oh yeah, again if you wanted a listing, that was another
pass.

Small wonder that our field service people would usually key in patches by
hand through the front panel, and just dump memory to tape to create a new
load tape.  If you were lucky, they would correctly record the patch by hand
on a paper listing (but of course finding the right listing could be
interesting as well).

Definitely the bad old days.  But solid state memory was just coming out,
along with more powerful machines and hard drives so things got better
quickly.

-- Randall


** triumphs at autox.team.net **

Donate: http://www.team.net/donate.html
Archive: http://www.team.net/archive
Forums: http://www.team.net/forums
Unsubscribe/Manage: http://autox.team.net/mailman/options/triumphs/dave1massey@cs.com

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://autox.team.net/pipermail/triumphs/attachments/20170419/e8437f38/attachment.html>


More information about the Triumphs mailing list