[ih] Fwd: History of "accounts"

[John Klensin]

On Sun, Feb 9, 2014 at 12:00 AM, John R. Levine <johnl at iecc.com> wrote:
>  Here's the manual for the slightly earlier Fortran Monitor System in 1961:
>
> http://www.computerhistory.org/collections/catalog/102663112
>
> On page 64 it says:
>
>   The first record of the Monitor is the "Sign-On" record. This
>   may be programmed by the installation to handle accounting or
>   other identifying information pertaining to a job.
>
> I expect that if we poked around more, we'd find more, earlier stuff.  In
> the 1950s computers were phenomenally expensive, and I find it hard to
> believe many of them were run without provision to charge back the costs to
> the users.  Unless there is some arcane kind of bookkeeping I never heard
> of, the way you do that is with accounts, maybe done in software, or more
> likely done with pen and paper.

Yes.  The "C" in CTSS stands for "Compatible" and what it was
compatible with was FMS batch and Timesharing running concurrently on
the same machine.   I also (vaguely) remember the 1401 instruction to
which Miles refers from the early to mid-1960s (but obviously going
back earlier).

Those decisions were, for the record, before my time -- I came to
those lines of machines a very significant few years later.

It would be interesting to check on whether explicit accounting stuff
appears in the line of machines and operating systems leading up to
the 707x -- one might imagine that chargeback regimes would have
appeared in "commercial" machines rather early.   But I don't remember
such arrangements on the 650 -- but have no idea whether that is
because they weren't there and charging was done by wall clock or
because my memory is gone.

Remember that "incredibly expensive" notwithstanding, accounting could
easily be done based on the wall clock (or even a punch clock, which I
vaguely remember in some installations) until operating systems got
smart enough to run more than one job at a time.

  -john (feeling very old today)