![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
The decline and fall of A/UX
Apr. 10th, 2021 12:29 pmThe story of why A/UX existed is simple but also strangely sad, IMHO.
Apple wanted to sell to the US military, who are a huge purchaser. At that time, the US military had a policy that they would not purchase any computers which were not POSIX compliant – i.e. they had to run some form of UNIX.
So, Apple did a UNIX for Macs. But Apple being what they are, they did it right – meaning they integrated MacOS into their Unix: it had a Mac GUI, making it the most visually-appealing UNIX of its time by far, and it could network with MacOSs and run (some) MacOS apps.
It was a superb piece of work, technically, but it was a box-ticking exercise: it allowed the military to buy Macs, but in fact, most of them ran MacOS and Mac apps.
For a while, the US Army hosted its web presence on classic MacOS. It wasn't super stable, but it was virtually unhackable: there is no shell to access remotely, however good your 'sploit. There's nothing there.
The irony and the sad thing is that A/UX never got ported to PowerPC. This is at least partly because of the way PowerPC MacOS was done: MacOS was still mostly 68K code and the whole OS ran under an emulator in a nanokernel running underneath it. This would have made A/UX-style interoperability, between a PowerPC-native A/UX and 68K-native MacOS, basically impossible without entirely rewriting MacOS in PowerPC code.
But around the same time that the last release of A/UX came out (3.1.1 in 1995), Apple was frantically scrabbling around for a new, next-gen OS to compete with Win95. If AU/X had run on then-modern – i.e. PowerPC- and PCI-based – Macs by that time, it would have been an obvious candidate. But it didn't and it couldn't.
So Apple spent a lot of time flailing around with Copland and Gershwin and Taligent and OpenDoc, wasted a lot of money, and in the end merged with NeXT.
The irony is that in today's world, spoiled with excellent development tools, everyone has forgotten that late-1980s and early-to-mid 1990s dev tools were awful: 1970s text-mode tools for writing graphical apps.
Apple acquired NeXT because it needed an OS, but what clinched the deal was the development tools (and the return of Jobs, of course.) NeXT had industry-leading dev tools. Doom was written on NeXTs. The WWW was written on NeXTs.
Apple had OS choices – modernise A/UX, or buy BeOS, or buy NeXT, or get bought and move to Solaris or something – but nobody else had Objective-C and Interface Builder, or the NeXT/Sun foundation classes, or anything like them.
The meta-irony being that if Apple had adapted A/UX, or failing that, had acquired Be for BeOS, it would be long dead by now, just a fading memory for middle-aged graphical designers. Without the dev tools, they'd never have got all the existing Mac developers on board, and never got all the cool new apps – no matter how snazzy the OS.
And we'd all be using Vista point 3 by now, and discussing how bad it was on Blackberries and clones...
Apple wanted to sell to the US military, who are a huge purchaser. At that time, the US military had a policy that they would not purchase any computers which were not POSIX compliant – i.e. they had to run some form of UNIX.
So, Apple did a UNIX for Macs. But Apple being what they are, they did it right – meaning they integrated MacOS into their Unix: it had a Mac GUI, making it the most visually-appealing UNIX of its time by far, and it could network with MacOSs and run (some) MacOS apps.
It was a superb piece of work, technically, but it was a box-ticking exercise: it allowed the military to buy Macs, but in fact, most of them ran MacOS and Mac apps.
For a while, the US Army hosted its web presence on classic MacOS. It wasn't super stable, but it was virtually unhackable: there is no shell to access remotely, however good your 'sploit. There's nothing there.
The irony and the sad thing is that A/UX never got ported to PowerPC. This is at least partly because of the way PowerPC MacOS was done: MacOS was still mostly 68K code and the whole OS ran under an emulator in a nanokernel running underneath it. This would have made A/UX-style interoperability, between a PowerPC-native A/UX and 68K-native MacOS, basically impossible without entirely rewriting MacOS in PowerPC code.
But around the same time that the last release of A/UX came out (3.1.1 in 1995), Apple was frantically scrabbling around for a new, next-gen OS to compete with Win95. If AU/X had run on then-modern – i.e. PowerPC- and PCI-based – Macs by that time, it would have been an obvious candidate. But it didn't and it couldn't.
So Apple spent a lot of time flailing around with Copland and Gershwin and Taligent and OpenDoc, wasted a lot of money, and in the end merged with NeXT.
The irony is that in today's world, spoiled with excellent development tools, everyone has forgotten that late-1980s and early-to-mid 1990s dev tools were awful: 1970s text-mode tools for writing graphical apps.
Apple acquired NeXT because it needed an OS, but what clinched the deal was the development tools (and the return of Jobs, of course.) NeXT had industry-leading dev tools. Doom was written on NeXTs. The WWW was written on NeXTs.
Apple had OS choices – modernise A/UX, or buy BeOS, or buy NeXT, or get bought and move to Solaris or something – but nobody else had Objective-C and Interface Builder, or the NeXT/Sun foundation classes, or anything like them.
The meta-irony being that if Apple had adapted A/UX, or failing that, had acquired Be for BeOS, it would be long dead by now, just a fading memory for middle-aged graphical designers. Without the dev tools, they'd never have got all the existing Mac developers on board, and never got all the cool new apps – no matter how snazzy the OS.
And we'd all be using Vista point 3 by now, and discussing how bad it was on Blackberries and clones...
