• Updated 2023-07-12: Hello, Guest! Welcome back, and be sure to check out this follow-up post about our outage a week or so ago.

Could A/UX Run On An 040?

Gorgonops

Moderator
Staff member
I had 72mb of RAM in my Q650, and even that is substantial overkill for A/UX. It was really quick and responsive on that machine; the MacOS emulation is based on System 7.0 and the UNIX underpinnings themselves are quite tiny and light by today's standards (reminiscent of Linux back in the 1.3/2.0 kernel days) so *for the software you can make run on it* I'd say A/UX runs better than MacOS 8.1 does on the same hardware.

If you skip the MacOS emulation and log into a plain X11 session it makes a passable X terminal if you have another UNIX machine on your network, but I *think* it only supports 256 color depth, so between that and only supporting X11R5 out of the box it's pretty hit and miss running "modern" software through it. Likewise the "seamless" MacX server that runs inside the System 7 sessions is a really fun toy, but it's fragile and pretty limited. Be sure to set your expectations appropriately.

 

Unknown_K

Well-known member
A/UX 3.x on a WGS 95 with the PDS card and full cache makes for the fastest 68k mac file server you can have. The built in Retrospect for A/UX is also very good for backup to DAT drives. A/UX was really used for workgroup servers and not really for workstation use (by Apple anyway).

 

NJRoadfan

Well-known member
I got the feeling that A/UX was a "solution in search of a problem". Its an artifact of the time when Apple was plowing money into various dead end projects AND an era where everyone wanted their machines to be a UNIX workstation. Heck the Amiga and Atari TT even followed suit. Apparently all these companies saw dollar signs in the UNIX workstation market at the time. Apple tried to recoup something out of it by positioning it as a dedicated server OS with specialized hardware (which went nowhere).

 

Gorgonops

Moderator
Staff member
Re "solution in search of a problem", you could certainly make an argument that the major software companies like Apple and Microsoft were wasting their time reinventing the wheel struggling to produce decent multitasking operating systems (and repeatedly failing) when UNIX was right there all along. The thing that really killed UNIX in the PC space was AT&T's license fees. No normal user wants their computer to require an operating system that regularly costs thousands of dollars a seat and no software company wants to give the bulk of their profits to a monolithic competitor, particularly when most of the good parts were being written for free by university students.

If the court battle in the early 1990 that divorced the open-source BSD components from AT&T's control had happened a decade earlier (Perhaps Berkeley could have preemptively sued AT&T in 1983 to prevent UNIX from being identified as AT&T's property to sell when they were released from their consent degree; they could have made a decent case.) there's a *very* good chance that instead of the painfully reverse-engineered VMS clone we call "NT" that resides in the core of Windows there would instead be some highly modified version of BSD 4.3. And A/UX, or at least something based on a UNIX kernel, may well have become the mainstream line of Mac OS evolution, saving us all a decade of pain and floundering.

In any case, didn't happen that way. The A/UX that actually exists arguably is sort of a half-arsed product; it cost way, way too much for a mortal user and if you just need a better server then all that cute integration between the Mac Finder environment and the UNIX backend is pretty much wasted. In that respect, yes, it's a product that doesn't really make sense. Nonetheless it's still, uhm, "cute".

 

NJRoadfan

Well-known member
Re "solution in search of a problem", you could certainly make an argument that the major software companies like Apple and Microsoft were wasting their time reinventing the wheel struggling to produce decent multitasking operating systems (and repeatedly failing) when UNIX was right there all along.
Microsoft's struggles were mostly due to fighting with IBM with OS/2 and later going their own way. Unfortunately the PC platform was so entrenched with DOS that anything new had to accommodate its quirks along with the original limitations of the IBM PC 5150 (640k or die). Apple's own problems came from people who were able to write up big and ambitious ideas on paper, but not deliver an actual product (Copland and Pink). The IBM factor was there too... remember Talgient? MS revisited this situation with Cairo and later Longhorn's "3 pillars" (I'm still waiting for WinFS). Lets not forgot GNU Hurd and its micro kernel/servers concepts.

there's a *very* good chance that instead of the painfully reverse-engineered VMS clone we call "NT" that resides in the core of Windows there would instead be some highly modified version of BSD 4.3.
MS was going their own way no matter what. The mature Windows NT wasn't that bad of a product from a technical point of view. The security issues stemmed from its backwards support of applications expecting to run on single user OSes and later assuming everyone had administrative privileges. MS finally fixed that issue with UAC in Vista. Rule #1 of any *NIX is never run as root.

And A/UX, or at least something based on a UNIX kernel, may well have become the mainstream line of Mac OS evolution, saving us all a decade of pain and floundering.
While its based on BSD, this eventually happened with MacOS X. In the end, Linux landed up killing the market for AT&T's licensed UNIX on microcomputers/workstations.

 

Gorgonops

Moderator
Staff member
Microsoft's struggles were mostly due to fighting with IBM with OS/2 and later going their own way. Unfortunately the PC platform was so entrenched with DOS that anything new had to accommodate its quirks along with the original limitations of the IBM PC 5150 (640k or die). Apple's own problems came from people who were able to write up big and ambitious ideas on paper, but not deliver an actual product (Copland and Pink). The IBM factor was there too... remember Talgient? MS revisited this situation with Cairo and later Longhorn's "3 pillars" (I'm still waiting for WinFS). Lets not forgot GNU Hurd and its micro kernel/servers concepts.
Aaaand all of these are what I was referring to when I said "a decade of pain and floundering".

MS was going their own way no matter what. The mature Windows NT wasn't that bad of a product from a technical point of view...
Regardless of how good or bad NT is now Microsoft had their own version of Unix (Xenix) that ran in some form or another on nearly every early 16-bit CPU (PDP-11, Z-8000, 68000, x86) and largely lost interest in it because (in roughly equal parts) their involvement with IBM (which both rightly hated AT&T and had a bad case of NIH syndrome) and AT&T's own actions regarding licensing. Certainly *if* Microsoft had started with BSD as the core of Windows they would have commenced to make their version of it private and proprietary in a multitude of ways, possibly to the point that you wouldn't even call it "Unix" anymore, but the point is that instead of the first version of "NT", IE, a version of Windows that wasn't a crippled product depending on cooperative multitasking and MS-DOS, premiering in 1993 (and not going mainstream until Windows XP in 2001) they might have had something of roughly equal technical sophistication half a decade earlier.

Linux landed up killing the market for AT&T's licensed UNIX on microcomputers/workstations.
Poetic justice?

 
Top