[SunRescue] Offerings

[Sheldon T. Hall]

The machine with the paper tape, magnetic cards, real core, and vacuum 
tubes was an NCR 315 "mainframe."  The "console printer" was an IBM 
Selectric typewriter with its actuators wired into the desk-sized 
operator's station.  The rest of the design followed that lead.

Loads of blinky-lights, too.

The machine itself had a paper-tape reader, the desk/console, the CPU 
cabinet, seven memory cabinets, a printer the size of a Volkswagen, and 
four CRAM units.

CRAM stood for "Card Random Access Memory" - rack-sized machines, each 
holding 255 notched magnetic cards which the CRAM unit could select by the 
pattern of the notches, dropping them into a mechanism which could read or 
write the data on the cards.  Saying it was "Rube Goldberg" would 
compliment its inventors unduly.

Of course, the storage devices being called CRAMs, and the length of a 
routine name being limited to six characters, the cannonical name for a 
program's I/O routine was "CRAMIT."

I started as a computer operator in that shop, circa 1972; I was their 
chief programmer less than 18 months later.  The programming language was a 
macro assembler called "NEAT."  Which it wasn't.  NEAT embodied all the 
evils against which the modern programmer/philosopher rails: 
 self-modifying code, computed GOTOs, no data typing, no re-usable objects, 
etc.  "GOTOs Considered Harmful" wouldn't cover the half of it.

On the other hand, it was fun.  The CPU was the size of two refrigerators, 
and put out so much EMF that a well-constructed program could not only 
balance the bank's books, it could play recogniseable songs through an AM 
radio on top of the cabinet while doing so.

Since any part of a program could be code or data, depending on whether you 
were executing it or reading it, it was possible, nay, desirable, to have 
pages of very non-obvious stuff in your programs.  Like Sun's start-up 
scripts.

In my case, I perfected a program which not only ran the 1099 tax reporting 
forms, but printed out a "Snoopy" calendar for the new year if it was run 
between Christmas and New Years.  This meant that the calendar wasn't 
present on the test printouts when the program was run before Christmas, 
but the skeleton crew got 'em when they ran the real thing after close of 
business on the 28th.

I took a holiday from the computer biz after leaving the bank, only getting 
back into it in the late 70s.  I'll save the stories of those hacks for 
another time.

-Shel