Computers In The Dark Ages
Early Days At Apricot Computers
When I was 14 my maths teacher took our class to the local technical college as an introduction to computers. The college had a PDP8 computer from DEC and an array of Teletypes (electric typewriters which would print out on to rolls of paper).
He wrote a BASIC program:
10 INPUT “TYPE IN A NUMBER” ; X
20 PRINT “THE NUMBER AFTER THE ONE YOU TYPED IS” ; X+1
30 GOTO 10
It is rather a basic (forgive me) program. It would go on continually until you typed control-C. Remember this is printing out onto paper. There was not a screen (a VDU) in sight. Those of you with some understanding of BASIC (and this is the original Kemeny and Kurtz variety, not the “Visual” kind) will understand how simple this is.
I was hooked. I knew I wanted to work on, in, with, computers.
Eight years and a computer science degree later in 1985, I arrived at 27 Highfield Road in Birmingham, the R&D department of Apricot Computers plc. Absent mindedly whistling as I walked in I was cut down by the security guard who joked to me “Don’t whistle here, they’ll think you’re enjoying yourself”. On this basis I had plenty of occasions later to whistle.
R&D was a 2 story building. I was in the restricted hardware and software office on the ground floor. In here were the hardware guys that developed the first PCs and the software team I would be part of. My first job, and what was to become a recurrent one, was the integral modem. Apricot had a very successful email package called Micromail, this used the PCs integral modem to dial out and seamlessly fetch email from a BT central computer, hosted by a service called Telecom Gold.
Elsewhere in the small department, (of around 15) engineers were working on BIOS, networking, Windows (version 1), porting Xenix (Linux to those of you under 25, Unix to those of you over 45). We weren’t networked together; code was shared using floppy disks. Backups were done individually. Often printouts of code were taken to work on and review. These being done on large green ‘music-ruled’ paper.
Those who were lucky (I was) had a 10M hard disk (still called ‘Winchesters’ then because ‘hard disks’ were those multi-platter things the DP department had on their mainframes). The PC du jour was an Apricot Xi, a hard disk based machine based on the Intel 8086. All work was done on these.
There was a friendly rivalry between the software and hardware departments. The hardware guys would design all this nifty hardware really quickly, but then it’d take us software engineers (or ‘systems programmers’ in the day) a while to get it working, but then we’d have to debug the equipment and supply the magic that made it work (or so we held). We had access to alpha and beta code from Microsoft and early chips from Intel, indeed the man from Microsoft would come up every couple of weeks to see us.
The atmosphere was great, Here we were playing with software no one else had, on hardware no one else had. All effort was focussed on the next launch, either to an audience of invited dealers or at an exhibition.
One of the things that set Apricot apart was the working environment. We didn’t have pool tables in reception (This was Birmingham after all) but the place was open 24 hours a day. People came in the early hours and worked late. When I worked on a multi-processor in the later years I’d come in at 7:30 work to 5, go home for an hour then go back and work till 9. Microsoft, being in Seattle, was 8 hours behind us so to communicate with the team over there this was the obvious thing to do. Also working on Saturdays was quite common (and not paid).
As one did in the olden days, Friday lunchtimes were spent down at the Hagley Duck, a pub on the main road into the city that was annually drunk dry alternately by the students at the local Poly and University. Lunchtimes were spent eating at ones desk an unfeasibly cheap meal prepared by Maureen the cook. In charge of us all was Dr Peter Horne, who had come from CERN. Like all of us at Apricot he was young for his position. At one point we worked out 95% of staff [in R&D] were younger than 30.
The next big project was Candy Floss, what was to be the Apricot Xen. When the ‘plastics’ arrived we were really impressed. The previous machine had been the less than exciting F10, - the most remarkable thing about that being that it had a colour screen but that was not seen as important then. During the Xen project we perfected the networking system we were working on (1Mb/s) and built a whole range of diskless workstations to hang off the main server.
During this time we managed a move to a purpose built building at the University (of Birmingham) Science Park, this was at least twice the size of Highfield Road and gave us much needed space.
In utter secrecy the Xen-i was also being prepared. Seen as something of a sell out, but a needed one, the Xen-i was the first IBM-compatible PC we did. It was launched less than a year after the Xen. Critically it used a different expansion bus to the in-house architecture so a lot of drivers had to cope with both ways of operating.
At this time key technology had started to change. The Xen and Xen-I both used the Intel 80286. This was a different generation of IC to that which had been used in all the earlier PCs and was a generation on from the IC still used in the IBM PC. This allowed better graphics, more GUI based software, and was destined to be more of a business machine. This also had a side effect that allowed it to be switched from one process to another. Called triple-faulting this was described as ‘switching off the car to change gear’ by someone at Microsoft, but it worked. It allowed such technology as “Advance286DOS” to be developed, (this was later called OS/2.), and later for Windows/NT and most Windows variants thereafter.
The biggest change eventually was the Intel 80386. It’s fair to say that the architecture of all PCs are based on this pioneering chip, which allowed for the gear-change without stopping the car. It is this feature which allows Windows and Linux PCs to switch between one application to another and run virtual PCs, CMD windows etc.
Apricot developed the Xen-i386 to use this chip, and it supported the full range of software available at the time. MS-DOS, Windows 3.x and /386. Unix and Xenix, and OS/2
The next step was a move to IBM’s proprietary Micro-Channel Architecture (MCA). More sophisticated than the simple AT bus, and the idea was that we would second source IBM. Here apart from Big Blue, Apricot was pretty much on its own. This was a path it trod for a year or two before giving in to the AT architecture in much the guise that it is in today.
Software evolved from DOS to Windows, From Windows to OS/2 and /NT and then Windows 95, which we saw much earlier than 1995. Hardware became more compatible, less eccentric and was based on the evolving design of the Pentium [Why is the Pentium called the Pentium? well Intel, moving from the 80186, 80286, 80386 and i486 got stuck when they got to ‘586’ they had already used that name on an older product].
At the end of my time there we had multi-processor, fault tolerant servers, multi-media PCs supporting TV and music, all networked together but it was all very much ‘wintel’ in flavour. We had become compatible.
At some point whilst visiting the Santa Cruz Operation (now a Linux distributor, but then the vendor of its own flavour of Unix) we were shown a program running under Unix which when we typed in a very long and unfathomable string (starting with ‘http’ would display from a computer in CERN a menu with a list of computer companies, when you clicked on the names the screen then showed more information about that manufacturer. In short the Web
Mitsubishi Electric bought Apricot from its owner ACT in the early 1990’s. ‘Melco’ reorganised our factory in Glenrothers, Fife. Later on we worked with engineers in Japan on large computer projects as Apricot moved more into business computing away from general PC market. I left in 1996 a few years before Melco sadly closed the business.