The Univac Solid State Computer 90, Daily practice.
These days we are used to the fact that our hard-discs can contain thousands of files with programs, plug-ins, pictures,
banners, videos, history, settings and occasionally simple data in well defined databases, accessed by accounting-programs.
All these files are directly accessible, however we don't know when or were they were loaded. It just happens and it only takes seconds.
In the 1960s datafiles punched in cards were fragile and rather complicated to manage.
Punched cards contained programs, standard routines or alpha-numeric data.
A stack of ten centimeters high contained: 500 (cards) x 90 (columns) = 45 K Bytes of data.
Our USSC was used to manage the administration of an international company with thousands of employees.
For this task the stream of data had to be well organized.
Theoretically the USSC could read 36000 cards or 3.24 M Bytes per hour.
Card-decks could only be read sequentially and there was not an index on their contents.
Therefor to for running each of the many programs,
datafiles first had to be sorted in a new order or to be merged together,
Because each program required the data in a specific sequence.
So for each run of a program:
- the program had to be loaded from cards,
- the cards with data had to be sorted into a new order
- and had to be fed into the cardreader.
The magnetic-drum-memory was far too small to store more then one program and
too small to store data-files to be used by differend programs.
Operators had to collect the punched cards with actual data and prepare the stacks of cards to have them processed.
All program-runs and other operations were well scheduled.
In case the system, the computer and the people, had to be halted for incorrect data, this was costly
and it caused problems in working out the scheduled jobs.
In addition hardware-failures also did cost lots of time.
Since there was no secured data-integrity, all programs produced printed lists of the data they processed.
These lists had to be stored till after the end of the financial year.
Typing and reading errors, missing cards and even program-failures always occured,
therefor the possibility of searching back the data of all events in the computating-history was necessary.
It could take days or weeks for accountants or programmers to find their way through these piles of lists.
Each day easily hundreds of pages were printed, maybe 100,000 pages a year, a pile of about 10 meters of paper!
All this paper had to be stored since we possessed no back-up device like tape or a Randex storage unit.
Storing punch-cards for longer periods was full of risks.
Older cards could easily got damaged in the card-reading equipment.