Bring back Blinkenlights!!! 3275

Here's a GIF mock-up of the Blinkenlights on the front panel of a GEC 4080 minicomputer (rather small I'm afraid).

The two rows of lights enabled two of the CPU registers to be displayed. The pair of rotary switches next to each row are used to select which CPU register each row displays. Top row can display registers up to 20 bits, and the next row does registers up to 16 bits. (It was a 16 bit machine, but some registers could be combined into 32 and 64 bits.) The third row are toggle switches and were used for writing values into registers (not clear in picture as they are all in the up (binary 0) position). The current value on them was always readable as the 'KEYS' register to programs too. Fourth row of switches-lights were for controlling the internal I-O channel, and the rotary switches for setting various CPU debugging modes (single instruction step, single microinstruction step, halt on error, etc). Bottom row of switches-lights are for running, stopping single-stepping, and resetting the CPU, IPL, keyswitch, and mains power.

Oh, people actually got rather good at judging system health by just a glimpse of the front panel lamps. For example, if the system was stuck in a tight CPU loop somewhere, you would tend to get an unvarying fixed pattern. If the system was idle, you got a fixed pattern from the idle loop, interspersed a few times a second by interference from things which ran on regular timer ticks. If the system was busy with applications, the light pattern would dance around all over the place. Obviously, this would depend which registers were currently being displayed too.

If you actually wanted to use the front panel for debugging (and I did that on a number of occasions), you first had to go through all the bulbs, working out which ones were blown and replacing them (they were tiny filament bulbs). Since the drawer of spares always seemed to be empty, that usually meant stealing them from the lamps you didn't need (the I-O channel ones were rarely of any interest during debugging), or stealing them from another system in the computer room. Also, they were difficult to get in and out. Occasionally the glbutt would break, and if the lead-in wires shorted, you blew the 75 series driver IC which would then need replacing too.

-- Andrew Gabriel