Editing IBM System/3 (section)

==Problems with System/3==
The System/3 had no provision for halting a process once it had started to run. For instance, if a compile failed because of an error on the very first page, the user had to wait for a sometimes voluminous compile listing to print in its entirety. Users learned to reach under the printer and jostle the paper discharge chute, which would cause the machine to halt with a "P3" (printer error) displayed. The user could then dial in the response code FF to abort the run. Another way of stopping it was simply to press the green "Start" button on the console, causing the system to reboot.

Error codes were displayed on a two-digit [[seven-segment display]] (one of the first seen, and built with lamps rather than LEDs). The range of error codes included not only decimal and hexadecimal digits (as seven-segment displays are commonly used) but also a limited set of other letters; for example, "P3" was one of several printer error codes. A thick manual that came with the System/3 aided the operator in interpreting the error codes and suggested recovery procedures. The System/3 had no audible warning device, so a program that was not printing, reading cards, or causing other obvious activity could halt and the operator would not know it unless they happened to look at the status display. Models with the Dual Program Feature had two separate status displays.

Most/many users did not buy a console. Instead OCL code was either suppressed entirely or printed on the 5203 printer. The console offered by IBM slowed down program execution tremendously when it printed OCL commands, as it was basically a [[selectric typewriter]].

The concept of keying your punched cards through the console was a marketing ploy. In reality, the System/3 could not be a computer and a keypunch at the same time, so when it was a keypunch, no computing was possible. The original IBM System/3, which was shown in July 1969, had the keypunch console so they could offer a computer for under $2,000/month. In reality it was unworkable and almost invariably users acquired a stand-alone keypunch/verifier.

Later several OEM companies built 96-column keypunches, sorters, and collators. This took the 'heavy lifting' off of the MFCU and freed the System/3 for actual computing functions.

Most experienced System/3 users minimized use of the MFCU as much as possible, since it was a system bottleneck.