Part 1: Some Early Mainframes

When I was about 10 or 11, I had a friend whose parents owned a milk bar in Ultimo. Their shop was opposite the old Sydney Museum of Applied Arts and Sciences which was in the building on Harris Street which is now the Sydney Institute of TAFE.

Entry to the Museum was free. It was full of working models and it made a great playground.

One of their feature exhibits was an "electronic brain"that played noughts and crosses. It wasn'tt really a general-purpose computer but a hard-wired device.

Nevertheless, I'm not the only one who thinks that wanting to understand how that machine always won may have sparked their interest in programming.

The "giant electronic brains" at that time used valves but, by the time I got to uni, valves had been replaced by transistors.

At uni, there were no undergraduate course in computing but I did get to do a little programming in a simplified version of Lisp on an English Electric KDF9./p>

The KDF9 was one of the first computers with all solid-state logic, having about 20,000 transistors.  Its memory consisted of 32K 48-bit "words". Each bit of memory was made of little magnetic rings or "cores" held, about a millimetre apart, in a grid with wires. Bits of data were stored by changing the direction of the magnetic field.

My first real programming training was a Cobol course on a Honeywell 8200 which was among the first generation of commercial computers built using integrated circuits, rather than individual transistors.

TThe H8200 had a memory of 256K characters; a "character" contained 6 data bits. (The world didn't standardise on "bytes" until the late seventies.)

And it was a big computer in more senses than one. It was capable of running up eight programs simultaneously (using time-sharing) and needed a minimum of 550 square metres of air-conditioned space. That's it in the header.

Disk drive for Honeywell H8200

Disk drive for Honeywell H8200

Our installation had some of the first disk drives in Australia. These washing machine size drives took packs of eleven disks with ten writeable surfaces and had a total capacity of 35M characters.

My next round of formal training was in Fortran, some more Cobol and an abominable report-generation language called Nicol on an ICL 1904.

BBy this time, disk drives were standard equipment and the 1904 had 8M character removable drives. It also had a unique tape cassette system - not the little compact cassettes used in audio systems, but the big 10.5" reels housed in cartridges.

My first programming job was writing Cobol and some assembler (in a language called Compass) on Control Data Corporation 3300s. The CDC 3300 had 512K 6-bit characters of core memory could achieve a million instructions a second.

By this time, disks were replacing tape, although (as you can see in the photo) there was still plenty of tape processing. The disk drives handled removable packs with a capacity of 110M 6-bit characters.

BHP Newcastle computer room

Dual CDC 3300s in the BHP Newcastle computer room

Most input, including all programming was still on punched cards. But we were developing some of the very first commercial systems in Australia where the users had keyboards and screens.

In the beginning, we were working with the manufacturer trialling different keyboard layouts to try to decide whether they should be qwerty, dvorak or alphabetical.

IIn my next job, programming a Burroughs 5500, we programmers got to use the keyboards and screens - although, at first, the screens were 40-column LCD screens, like stretched calculator screens.

After the B5500, came a Burroughs 6700. Then I met my first microcomputers and that changed everything ... but that's another chapter.