Forgotten Voice

My computing experience prior to IBM had mainly been on a Hewlett Packard mini-computer used for laboratory testing by a loudspeaker manufacturer. The IBM technology was amazing, but its sheer scale and complexity was rather more difficult to assimilate. As a young engineer I was given a mainframe Virtual Machine with which to do my work, hosted by a mid-range IBM System/370 Model 145 in the air conditioned machine room in X block. The algorithms for sharing expensive computer resources worked extremely well and were in advance of anything running on Intel platforms well into the 21st Century. Multiple parallel work queues were dispatched according to workload characteristics with more interactive work taking priority over longer running jobs. While compilations and builds may take minutes, the shared mainframe dealt with nearly all interactions within a couple seconds and many within a second, although occasionally things ground to a halt. Forty years later and with vastly greater resources, the World Wide Wait often struggles to provide the same level of responsiveness as those early machines.

The VM was connected to IBM’s precursor to the Internet and in 1977, I was able to send and receive messages and files and form working relationships with IBMers across the globe. By the time I started travelling on business, I was able to log onto my VM from terminals in other countries across the network. Utilities could be shared on the equivalent of bulletin boards. Many young engineers and programmers had their own pet projects to contribute utilities to the pool. One notable example was the REXX programming language contributed by Mike Cowlishaw, which now has a life beyond IBM.

The engineers in X block shared about a dozen “green screen” IBM 3270 terminals which led to a different work pattern to the one that emerged a few years later when everyone had their own computer terminal. First we researched the problem we were set to solve and made design diagrams, pseudo code and notes on paper. Only when we were confident that we had a realistic approach to solve a problem did we start entering and testing code. In many ways, it was a good discipline, but it relied upon the availability of a comprehensive stack of paper manuals and reference cards and their acquisition, storage and use was sometimes a challenge. Obtaining a manual could delay a project by weeks. Some manuals were considered so commercially sensitive that they were printed on candy striped paper and stored in special, double locked filing cabinets with each access logged. Early computer operating systems provided no menus and few prompts. Messages were terse and often in a code which required reference to a manual. Filenames and file types were each restricted to eight characters. We became experts in unconsciously typing long strings of commands with their options and parameters. We dreamed of the day in which computers would auto-complete text, not that we knew anyone who worked on that idea. Even today, I will unconsciously type frequently-used words and phrases rather than the intended text.

The only graphics available to us were rectangular line drawings created from special characters loaded into the display’s character set. None the less, we created effective engineering diagrams that could be printed on continuous fan fold paper by a brute of a printer in the air conditioned computer room. The printer operated by spinning a long steel band on which the characters were embossed. Hammers forced the paper against the type at the correct instant to produce the required text. In effect, the printer was a high speed, high precision chain saw. It was so noisy that conversation was not possible anywhere nearby. While I was working in X Block, a few Hursley-developed IBM 3279 colour displays were installed which could display more advanced business graphics generated by the Hursley developed IBM GDDM software. These were exciting times!