🚀 go-pugleaf

From article <1992Apr15.152445.5780@sctc.com>,
by smith@sctc.com (Rick Smith):
> 
> About the only "first" historians seem to agree on is that the
> Manchester Mark I ran the "first stored program" of any computer.


"What was the first programmable computer?" is ill specified question.
The first machine to be completed is not necessarily the first one to
have run a stored program, because it is pretty clear that the early
machines were working programmable computers for a significant time
before they were debugged and fully operational, and it is also hard to
tell from the records when the first program was loaded and tried for
debugging purposes.

Back when I was a student at Illinois, I started reading back through
the old Digital Computer Lab technical reports, wondering about the date
of the first program run on ILLIAC.  The old reports from the late 1940's
were mostly dittoed, and much of the ditto ink had turned into faint
brown shadows, making these reports very hard to read.

What I found was that ILLIAC I and its twin ORDVAC came up slowly.  Long
before there were any serious reports of any applications programs being
run, there were pages of teletype output (judging by the type font)
stapled to the dittoed reports.  These were octal dumps of the machine's
memory, included as supplements to the tech reports to demonstrate that
the Williams tube memory was finally working.

To make such dumps, there must have been a considerable amount of working
hardware -- not an entire working computer, but enough to step throug
memory and convert binary to octal for printing.  My suspicion is that
they used the CPU's data paths to do this, with a very simplified control
unit that turned what they had into a memory tester, but I didn't have
time to find out if this was so.

I also got the impression that they were running programs with just add,
jump, load and store instructions for a considerable time while the
remaining parts of the control unit were debugged enough to announce that
they'd built the machine that they'd contracted to build.

The date used for "official" purposes for when Illinois got its first
computer running was the date on which ORDVAC passed its acceptance tests
(after which it was crated up and shipped out).  There was an interesting
side note to this:  ILLIAC and ORDVAC were built at the same time, but
there was apparently no binding of a name to a machine until one of them
passed its acceptance test.  At that point, the working machine became
ORDVAC and was crated and shipped, and the other one became ILLIAC and was
finished for local use.
					Doug Jones
					jones@cs.uiowa.edu