That's a great story Kevin! That'd be a photo I'd be interested in seeing,
of those cards (or if they're in some museum archives somewhere). Or in
some abandoned IBM warehouse.
I'm just trying to envision the workflow (since certainly I "wasn't
there"
in the 60's). There is a diagram in early editions of Operating Systems
Concepts, that shows a deck of punch cards - showing they contain CODE and
DATA. You'd have to load the FORTRAN compiler itself, insert a control
card to invoke the compiler, and then another control card to RUN your
compiled code. Translating that to modern workflow, I view it as like
loading the "runtime library" of your compiled code (the "little
things"
like math across multiple words, or interacting with whatever I/O devices
were available).
Also thanks Donald, neat link on punch cards, I hadn't come across that.
So whose got a similar thing for punched tape? (can we call it "fan tape"
or is that a separate category altogether?)
-Steve
On Sun, Mar 1, 2026 at 2:04 PM Kevin McQuiggin <mcquiggi(a)sfu.ca> wrote:
Hi All:
My comments, worth two cents or less.
The first computer I ever used was an IBM 1130 my city’s school district
office at when I was a 14-year-old high school student. This was in 1974.
My exploration of the “IBM Disk Monitor System" on the machine led to me
clobbering the FORTRAN compiler on the removable hard disk by mistake one
evening. The disk had a form factor similar to the DEC RK05.
Staff discovered the problem the next morning and had to reload the
compiler from punched cards. They told me that they restored the entire
system disk (OS and FORTRAN) rather than just the compiler. I saw the ~5
drawers of punched cards that they used for this. I was mortified at my
mistake and expected that I was in BIG trouble. Apparently the restore
took most of the day.
In this case it looks like the OS was held as backup on cards. It was a
small installation and there were no peripherals other than a card reader
and the integrated system console.
In bigger installations I agree, magtape was probably the standard media.
Postscript: don’t submit a DUP job:
// XEQ DUP
*DEFINE VOID FORTRAN
on an 1130. This does NOT give you examples of invalid FORTRAN code or
errors, as my young mind surmised, but deletes the FORTRAN compiler.
I got in trouble, but it wasn’t serious as I made an honest error, driven
by curiosity.
Kevin
On Mar 1, 2026, at 11:47 AM, Paul Koning via
cctalk <
cctalk(a)classiccmp.org> wrote:
> On Mar 1, 2026, at 2:34 AM, Steve Lewis via cctalk <
cctalk(a)classiccmp.org> wrote:
...
Anyway, apologies - it was just something that only recently occurred to
me, that basically all of the original operating systems originated on
punch cards: CTSS, Supervisor, AOSP, SCOPE, even MULTICs.
That's certainly not true. It may be true for card-centric outfits like
IBM. I don't know what early DEC development looked like, but considering
the rarity of card handling equipment on DEC systems I would expect paper
tape.
Early software for the Dutch machines I know was done on paper tape. In
some cases that involved punch equipment with custom-designed coding; for
example, the Electrologica X1 had a rudimentary assembler in ROM (along
with a BIOS) and source text was given to it on 5-channel paper tape, in a
code slightly above straight binary machine language.
Its successor the X8 had paper tape I/O standard, and the standard
executable file loaders used paper tapes. Ditto the bootstrap. The famous
THE operating system was a paper tape batch system, with the OS image
supplied on tape (though I think at some point it was moved to magtape for
faster startup). No punched cards were seen there until the X8 was
replaced by a Burroughs 6800, circa 1974, and even that machine had paper
tape input to support all the applications that had the input data on paper
tape. My father's precision measurement lab (part of the ME department)
had instruments that punched the measurements onto paper tape, for later
processing by that central computer system. All that was in ALGOL, by the
way.
paul