In 1984 (ish) Paul Berger and I set up the Australian company Applix (no relation to the US company) to commercialise the Applix Plus80. The Plus80 was a little Z80 computer which, when connected to an Apple //c allowed the user to run CP/M80. CP/M ran entirely on the Z80, using the Apple for display, keyboard and disk storage. We sold literally tens of these systems.The 1616 project history
Fortunately we had bought a few thousand Apple II Z80 cards ("StarCards") from a company who had bought them simply to unbundle WordStar. These cards were onsold to dealers (sans WordStar) at a tidy profit.
Flushed with the success of the Plus80, Paul thought it'd be cool if "we" (as in "Andrew") were to develop a 68000-based personal computer for sale in kit form.
I couldn't see a good reason to the contrary, so we traipsed into see Jon Fairall at the Australia and New Zealand Electronics Today International magazine and asked him if he'd be interested. He was.
Paul looked after the retail premises and kept on dreaming up new features (I called it "ed". He wanted "edit". It was called "edit").
I went to work. Every evening I came home from work and put together the 1616.
The prototype was built on a SpeedWire development card. Later the schematics were sent to Geoff Hart of Circuit Board Designs for layout.
The computer needed a "monitor program", or course. And an assembler. We licensed the source to a 68000 assembler from Peter Farb's company "Farbware".
The development system was a CP/M68k box running Digital Research's C compiler. That compiler didn't have unsigned longs, which sucked. Big time. We picked up an Apple Lisa running Xenix. It had a megabyte of RAM and a ten meg disk with a transfer rate of 10 kbyte/sec. Fast it wasn't, but it did have unsigned longs.
In December 1986 I published the first of four articles in ETI magazine. Systems started shipping. Lots of them. We dropped the 128 kbyte option; all systems had 512 kbytes. These ran 1616/OS v1.x. Mass storage was a RAM disk and audio cassette running at a blinding 3300 bits/sec.
In 1987 Kathy and I were interviewed by Gareth Powell, the doyen of the Sydney Morning Herald's travel and computer sections. We were the lead article on that Tuesday's IT section. Our 15 minutes.
In mid-1987 the Z80-based disk controller card started shipping. 1616/OS revved to 2.x to support it.
One of the first floppy-based systems was shipped to Clyde Smith-Stubbs at Hitech Software in Queensland. Clyde had gratiously agreed to port his C compiler to the 1616 platform. My idea of how file descriptors worked nearly drove him nuts, but he stuck with it, for which I am eternally grateful. I quickly moved the OS development to Hitech C, thus dumping the Lisa before it dumped me - we'd already bought another for spare parts.
The amazing Conal Walsh took the Z80 card and put the ZRDOS CP/M80 clone onto it. Once MGR came along it was possible to run CP/M within an MGR window!
In 1988 1616/OS v3.x was completed. It supported hierarchical file systems.
In 1988 (or was it '89?) the memory board started shipping. It had up to 4 megs of RAM, a discrete logic MMU and an NCR 5380 SCSI controller. (The SCSI controller on the Z80 disk card was too slow). 1616/OS revved to v4.x with the addition of preemptive multitasking and multiple users.
In 1990 (?) I ported the MGR window manager (Screenshot).
Around 1992 I ported version 2.5.8 of GCC and G++, binutils and GDB, thus proving that this version of GCC could compile itself, with optimisation, in 4 megs of RAM.
Andrew Morton, 8 March 1999