Hacker News Reader

Comment by jimjag | Hacker News Reader

I truly believe that A/UX doesn't get the credit it deserves. Was it ambitious? Sure. Did it it fully succeed in the goals it set for itself? Not really. But it was really a very nice integration of the Finder and UNIX. After all these years, I still look back on it with fondness. It certainly was a cornerstone platform for me.

View on HN

For those who don't know, Jim is/was the person who maintained the A/UX FAQ and often bailed people, like myself, out of corners when we got into them.

Thanks Jim!

Interesting that the article calls out the price point as one of the reasons it didn't succeed. Although a lot of that was the OS only having drivers for expensive hardware and being unusable on the lower end Macs.

That said, high price killed off most of the legacy Unix vendors/products. The world was changing around them and they weren't ready or able to give up the profit margins they had enjoyed for so long. Personal computers were becoming commodity equipment and free versions of Unix were hard to compete against.

It was, in retrospect, a mistake not to embrace good enough and, instead, keep pushing towards the high end of the market when good enough was what most users needed.

Of course a Unix workstation was twice as fast as a PC and cost twice as much, but it becomes pointless when most of the time it’s bored to death waiting for me to move the mouse.

The irony for me is that Unix was the (unhappy) inspiration for the famous “worse is better” memo. The Unix workstation vendors made high end hardware with (semi-) commodity software, and had their lunch eaten by generic hardware coupled with software that gave more people a better experience.

(Though personally not at all a fan of most of the Unix paradigm, for me it’s a vastly superior experience to Windows. But I can’t deny that that is not the case for most people)

“Worse” was Bell Labs: portable C/Unix. As opposed to MIT: codesigned Lisp/LispMs. The Unix workstation was the triumph of COTS micros over LSI or custom VLSI hardware.

> The Unix workstation was the triumph of COTS micros over LSI or custom VLSI hardware.

Well the Sun-1 definitely started there, no question (I don’t remember the Daisy or Apollo hardware). HP definitely never did and Sun (and SGI et al) all went down the custom hardware rabbit hole.

By the time they tried to hop onto the PC hardware train it was too late. None of those companies survive in any meaningful way.

BTW if you catch this in time to edit: you might want to put a hyphen between “co” and “design” because you didn’t mean signing code.

My impression is that they all built desktop minicomputers possible thanks to CPUs like the 68K but moved on to RISC designs when the 68K started showing its age. I would not say the PA-RISC was open, but SPARC had multiple sources and MIPS showed up everywhere. At that period, the x86 was not an option - Sun tried.

Apollo was building bit-slice 68000 emulations (to have an MMU) into the mid '80s, and ran Aegis, their homegrown fully-networked GUI OS, coded in their home-grown Pascal, with their home-grown touchpad pointer and home-grown token-ring network, on those and on actual 68K into the late '80s.

Aegis was inspired by MULTICS, not Unix, and was definitely a better system. They were demand-paging across the network in the early '80s.

One feature I recall stood out: they expanded environment variables in symbolic link text, like /usr/bin -> /usr/$SYSTEM/bin to get a SYSV or BSD Unix flavor, later on. The only Unix that does something similar I know of is Dragonfly.

> Aegis was inspired by MULTICS, not Unix, and was definitely a better system. They were demand-paging across the network in the early '80s.

I’d love to see one of those operating. We have lost so many great ideas we seem to never revisit…

Another was a read() system call that would copy into a caller-supplied buffer if it had to, but would normally just return a pointer into its buffer cache.

I know somebody who has some. You sort of need more than one so the token has somewhere to go.

The customness of the hardware is partly relative. One had to guess the trajectory of the PC to bet on clones and their components. Of course at one point there was no question the non-x86 workstation used "custom" hardware vs. the kind of more open ecosystem of x86 PC components, however even in this situation doing custom is not even an absolute criteria for success or failure or even eventual economy of scale: case in point Apple. Now of course there is in-house design vs. OTS but while it was true that the first PCs used pre-existing chips, quickly some chips started to be developed specifically for PCs or at least with PC as the main target, by far. So it is also kind of "custom", just developed by multiple companies.

Now in retrospect some workstation vendors could maybe have survived a little more by switching to x86 PC like hardware except the window for doing the switch was astonishingly small and they would have transformed to either a random OS vendor, or a random PC hardware vendor, or even both (even if requiring their own hardware, their competition would have quickly been way more directly e.g. Linux or BSD on generic PCs, and eventually with e.g. CAD vendors switching to Windows it would not have helped either)

Or as a random hardware PC vendor, what is even the point compared to their initial positioning and what was a "workstation". This market is now taken mostly by chip vendors with more or less artificial market segmentation -- and then computer vendors using such chips but they do not define the platforms anymore and add far less value. It's kind or logical; well at least in retrospect, here too. A very few number of platforms had to remain because of both the network effect and the practicality of using and developing for them. And consumer hardware was bound to eventually get state of the art designs (mostly scaled with parallelism for pro hw + a few artificial market seg)

You can take the internal dev route (again: Apple) but you had to target the general public first to do that (so not appropriate for a WS vendor)

Ironically, we could argue that to survive "in a meaningful way", if I read that in yielding a legacy today that could influence the ws workload by providing them at least a part of the platform, the old-school Workstation vendors would have needed to pivot to more pure component makers (for PCs).

Now in retrospect some workstation vendors could maybe have survived a little more by switching to x86 PC like hardware except the window for doing the switch was astonishingly small and they would have transformed to either a random OS vendor, or a random PC hardware vendor, or even both

I think this window was non-existent: Moore’s Law at the time was turning white boxes into workstations faster than any time-and-money consuming custom engineering could pay back the investment.

I'm not sure about that, even today. The typical white box PC motherboard then (and now) doesn't support 128gb of memory for example, you have to buy a HP Z400 or a Mac Pro (or an old server). Plus they aren't really designed to be maintained, or have space for dual processors. The Sun/HP/DEC/IBM workstations were not purely bought because they were fast, there was often specific software in mind that the end user wanted.

I have two primary machines I do development on, an Acer gaming laptop (a few years old, intel i7 8750h, 32gb of memory) and a truly ancient HP Proliant server with dual Xeons (x5650 and 96Gb of memory). The Proliant is still consistently faster at compiling an Android app than the laptop, despite the laptop having SSDs and the Proliant being 7 years older. There is a lot more to making a workstation than just raw CPU speed which is as true now as it was in the early 1990s when a Sparcstation was the go-to performance machine to have on your desk.

Same here. The 5yo Lenovo trounces the 2yo top of the line x86 MacBook Pro on pretty much everything. When it was young it easily humiliated every computer in the house.

the SUN-1 was based on the SUN (Stanford University Network) board, but I can't find anything concrete on the SUN board (Also by Andy) other than it existing, and apparently 'cheap/free' to license?

Apparently it formed the foundations of cisco/SGi & SUN.

Indeed, I remember those days well.

I was there at that time, there are manuals in white binders down the hall from me now. The Mac was made to never, ever, use the command line. That is what I chose to develop for at that time, including exposure to "high end" machines at University.

fun fact - there were no undergraduate courses in computer science at that time; master's level and up .. you have to be trained to use those workstations, even for five minutes, AND the oversight of an admin with security.

Mac? get one, fire it up, make Mac Paint pictures. The network IS NOT the computer, thankyouverymuch

fun fact - there were no undergraduate courses in computer science at that time

Perhaps this was regional. I was enrolled in undergraduate Computer Science courses at this time, and it was at a pretty low-end state university.

I actually wrote a large chunk of the A/UX unix port (late 80s) - a decade before (mid 70s) I'd obtained an undergraduate degree in Comp Sci (in New Zealand) - undergraduate Comp Sci was very much a thing at the time

not at Berkeley ! but proper respect to you Taniwha, many paths

Well I did the A/UX port IN Berkeley (the city, not the university).

The thing is that as a discipline Comp Sci is a late comer, university Comp Sci departments came from lots of places, some grew out of Engineering depts, others from Maths, Physics, Chemistry, Commerce, others from the computing infrastructure groups withing universities - they ended up being called all sorts of things - early on places offered Comp Sci by a whole lot of names

Yes, Berkeley had undergrad CS degrees in the 80's (and late 70's). One in the College of Engineering and one in the College of Letters and Science. Also, an undergrad EECS in Engineering.

The Bay Area school that didn't have an undergrad CS program was Stanfurd.

Back then MIT had one but Harvard did not. I felt sorry for my friends who opted to go there.

H recently started an engineering school. Years ago they tried to buy MIT but were rebuffed.

evidence welcome - I do not recall that as the case

Surprisingly tricky to prove. It looks like Berkeley overhauled its student body statistics about 15 years ago, and data from before seems to have vanished.

Is it argument from authority if I cite Karp?

https://www2.eecs.berkeley.edu/bears/CS_Anniversary/karp-tal...

The two undergraduate degree programs at Berkeley seem to date from 1968 or so. (Karp is fuzzy about his citations).

It was certainly well established when I went there in the early 80's.

Have you ever played with this:

https://macintoshgarden.org/apps/aux-apple-unix-68k-version-...

I found it interesting that it mentions both ‘Milwaukee‘, and Cayman "brac" ... Here on the Jasmine 80.

Seems like lots of weird codenames.

It looks like UniSoft had done a lot of work on the code (you?) to make it super portable. Although everything not directly related to the Milwaukee was cut. It'd be interesting to see if the kernel could be built without MMU support.. I tried to remove PAGING but that didn't work. Oh yeah the kernel source is on that image and other than one damaged file, it not only builds, but works on a special Shoebill.

cd /sys/psn

rm *.o

cd io

mv screen-data.c screen_data.c

cd ..

make unix

Pretty neat, none the less!

(here's probably too much information)

yup I wrote a lot of that, I did all the console stuff (font renderer, vt100, mouse, keyboard, fdb/adb, ui event queue etc).

Missing is all the auto config support (ability to add 3rd party drivers to the kernel in the field, something that at the time you usually couldn't do without source), fixes for large screen support, all my appletalk code is missing too. However I can see the latish floppy bug fix (there was a hardware bug found in the FDB/ADB chip where if we accessed the floppy on only some systems the keyboard froze, turned out we were spinning reading the timer while reading the disk, doing it so fast the clock to the FDB chip sped up) and the older pre-fix code in "sony.c.~8A" - that sort of places the release in time.

Given all that I'd guess that this is one of the early releases we gave to Apple that has somehow escaped - we used to fly to Oregon and copy disks to do releases to avoid California's sales taxes - UniSoft actually had an online modem based system where customers could log in and download the latest software - at the time CA tax law didn't explicitly cover electronic delivery of software, we were sued by the state and they lost setting a precedent

It's been a long time but "psn" probably stands for "Pigs in Space" - our internal code name for the project (Apple's code name "Eagle" leaked, but ours never did).

The "Cayman/brac" looks like what someone named their box, "Jasmine" was a type of hard drive, I've no idea about Milwaukee - I think Cayman used to be a company that made mac ethernet hardware, my guess would be that someone at Apple slipped them this pre-release, and then an update with some bug fixes, along with source (very naughty! AT&T would likely have sued us, ie UniSoft if they found out)

The assembly font renderer uses the 68020 only bfins bit field insert instructions, it can under certain circumstances generate a 24-bit bus write, nubus doesn't support such a thing, we got half of the first big run of mac 2 boards, they came with a schematic and PAL equations, I got to send fixed PAL equations back to Ron at Apple to make it work.

As far as "making it portable" a large part of that is just system 5 (SVR2 in this case) not so much us, though porting this code was our bread and butter - we supported a number of MMUs - 2 paging MMUs for 68020s, and a whole lot of swapping MMUs for 68010/68000s (plus 29k 88k etc) - it was never really designed to work without some form of MMU

Cayman must be Cayman Systems (Cambridge, MA).

I'd seen this zip file floating around for well over a decade, maybe quite a bit longer. It was amazing to see someone write enough glue for an emulator to actually boot it. It's even more crazy to get 3.0.1 to mount it under Qemu and do a full build in 17 seconds... I can't even imagine how long it took to build this back in the day!

Apparently above Milwaukee has something to do with Gasse and the BigMac Jr.

I wonder what ever happened to Unisoft's unix business. I've always wanted a SYSV but they seem impossible to buy. Best I have is a non commercial SYSIII from SCO before they 'gave 32v' and lower away but it's all so murkey if they could give anything away but I think they could sublicense for a fee (best $100 ever!)

After I left the Unix biz changed, we'd made money doing lots of ports for small companies, as Sun/SGI/etc got successful there were fewer of them, then we did the 88k and 29k ports but those guys didn't really take off. Doing A/UX also kind of screwed the company, I think it made us ignore our smaller customers, we expanded a bit too much, moved into a nicer building. We also didn't do an x86 port, which was a mistake - we actually helped SCO with the MMU portion of the SysV 286 port, there was some deal where we were supposed to get a copy but somehow that didn't happen (no love lost there).

UniSoft still exists, these days they sell digital video stuff (which was very weird when I became a cable/sat protocol engineer for a while a decade or so ago).

Reading what other people have been saying "Milwaukee" might have been a Mac2 code name.

I assume it can't have been that common not to have undergrad courses in cs?

I know my (non US) University have had undergrad courses since 1970.

> Unix was the (unhappy) inspiration for the famous “worse is better” memo.

It could be rephrased as “done is better than perfect”. I would love to have a high-end workstation based on exotic hardware with ridiculously fast storage, but an average home PC is probably enough for my development work and, when it’s not, I can acknowledge it is so because the software is much more bloated than it should be.

The Macintosh was the "good enough" version of the "Apple Lisa".

https://en.wikipedia.org/wiki/Apple_Lisa

It saddens me the beautiful stationery metaphor of the Lisa got ditched in favour of files and applications…

The Lisa deserved better.

That and Object Pascal being replaced by C++.

C++ was not yet released when the Mac was released.

Indeed, and what I am talking about happened after the Mac being released.

> By this point, in the late 1980s, the market was moving towards C++, and the beta version of Apple C++ compiler appeared in 1989, around the MacApp 2.0 release.[5] At the same time, Apple was deep in the effort to release System 7, which had a number of major new features. The decision was made to transition to an entirely new version of MacApp, 3.0, which would use C++ in place of Object Pascal. This move was subject to a long and heated debate between proponents of Object Pascal and C++ in the Usenet and other forums. Nevertheless, 3.0 managed to garner a reasonable following after its release in 1991, even though the developer suite, MPW, was growing outdated. Apple then downsized the entire developer tools group, leaving both MacApp and MPW understaffed.

https://en.wikipedia.org/wiki/MacApp

And then most just moved into Codewarrior and PowerPlant anyway.

If the workstations were twice as fast and cost twice as much they would still have a good market niche. Instead they were more like 30% faster and cost 10 times as much.

Intel eroded that speed delta pretty quickly too. The other chip makers were crushed by Intel's billions of R&D spending.

Motorola's floating point was 5x as fast as Intel's.

That wasn't enough to overcome market positioning.

more so the platform + the OS... I guess being in for 4k for a base machine (probably 2k more to make it usable) 500/1000 more isn't going to break it, but 386BSD machines while far less usable would be a heck of a lot cheaper.

It was expensive. In 1987 I built a 386-25 with 16MB RAM and SCSI HD that ran Interactive Systems UNIX for a lot less than it would have cost to get a Mac II with A/UX.

16MB in 1987, good lord. I didn't even know that was possible on x86 at the time.

16MB systems were still mainstream in 1997! Lots of 166Mhz and 200Mhz Pentium systems sold with 16MB of RAM that year.

When Windows NT first came out in 1993 its 12-16MB memory requirement was considered to be a major obstacle to adoption, IIRC.

That was using socketed DIL chips, 8MB on the motherboard and another 8MB on an expansion card. I had to insert the chips individually into the sockets. I had mis-remembered the disks, it used ESDI not SCSI.

Interactive still did beat SCO/Xenix on price in 1992 for running Oracle (on 386 hardware that costed as little as a single 857 MB hard disk on RS/6000 with AIX). And Novell 2 or 3 sure did beat A/UX as AppleTalk file server.

> That said, high price killed off most of the legacy Unix vendors/products.

High price and the differentiation. In the 90s I was working in publishing and publishing-adjacent companies that were throwing out their Unix environments as quick as they could after NT 4 was released. They were all sick of dealing with the crap of having multiple, subtly incompatible *ix environments to deal with: one company I worked for had Irix, AIX, SunOS, Solaris, and both flavours of Digital's Unix products, because various vendors had done deals with the different vendors to ship their products on those variants. Each came with different shells, different userlands, and all required slightly different tweaking and tuning to maintain operationally.

Contrast that with the NT 4 world, where that same business was quite happy to buy extremely expensive Alpha/NT systems so that the same skills, tool, and so on that worked everywhere.

Imagine in today’s money sinking 12,000 USD on a computer.

Do you take the Macintosh II with A/UX aka Unix from a company that doesn’t seem all too interested in the product themselves and selling what seems like a toy or do you go with SUN? They hired Bill Joy, and they are 100% committed to Unix?

They simply cost too much for a non commital company like Apple. If anything it’s amazing they saw it through to the end of the Quadra

Ah yes, Jagubox :-)

A/UX was easy to integrate into a mixed SunOS, Solaris, Irix network with liberal use of arch-dependent automounting.

Yes, when I went looking for all the bits and pieces, jagubox had them. Thanks Jim. I used this on a SE/30, a Turbo IIci, more than one Mac IIfxs.

I also thought it was pretty nifty little OS, courtesy of an A/UX office file server that I was tasked to set up in 1993 or so I think. And yes it was a very expensive file server at that.

My recollection was that internals of A/UX emulation were reused to enable the PowerPC transition from Motorola, but I might be wrong about that.

Unix licenses at the time were ridiculously expensive, even for PC-grade hardware. Adding compilers and development tools was frequently an extra eye-watering expense.

which to me was surprising about A/UX, you got both C89 and F77!

I think GCC was available back then. Not sure about F77. What C compiler did BSD use?

BSD 4.3 still used the Portable C Compiler. By 4.4 Berkeley was starting to use GCC, but that wasn't available in time to incorporate into A/UX. Plus, IIRC the GNU project was actively boycotting Apple at the time, and so Apple was probably adverse to incorporating any GNU software.

Out of the box running strings on stuff shows the 1984-1985 AT&T-IS 1985-1987 UniSoft Corporation strings.

So it's got to be PCC.

The F77 credits Apple, Adobe, AT&T-IS, Motorola, SUN, CSRG, and Unisoft.

In the A/UX 0.7 build there is a 'greenhills' marker in crt0.o and libc so it looks like they used greenhills before switching (self hosting?) to pcc?

+1 on pcc.

Greenhills was available as a third party compiler. Not sure why there would be markers in crt0 and libc, but perhaps someone at Apple rebuilt. Greenhills was better for most code that mattered.

Can't see crt0 mattering for performance, but maybe there was some interoperability glue to make both runtimes happy.

without the full source it's impossible to say. That said there was a 'tar ball' of 0.7 floating around for well over a decade, but it's able to run under emulation now.

https://macintoshgarden.org/apps/aux-apple-unix-68k-version-...

there is a /usr/lib/greenhillls with a ccom68 fcom68 and pcom68 along with some libs and crt files. It's all binaries but I did a compare on crt0 and they match the system one. Maybe it's just a case of the crt0 being assembly and it being.. well the same assembler.

Maybe they were just investigating PCC vs Greenhills or it was a cross thing, or they built PCC with Greenhills. who knows?!

We had 286's running Microport Unix at the company I was at back in the 1980s. I seem to recall that we were using Microport because it was pretty cheap. Compilers and all. I could be mis-remembering.

> My recollection was that internals of A/UX emulation were reused to enable the PowerPC transition from Motorola

No, pretty sure this wasn't the case. The RISC LC group used a custom emulator and nanokernel which is not at all similar to A/UX. The RLC couldn't even run A/UX, which was why Apple talked about porting it to OSF/1 for the new Power Macs (which, of course, never happened).

Right, from memory that was going to be A/UX 4.0. But as you say, that never happened.

> My recollection was that internals of A/UX emulation were reused to enable the PowerPC transition from Motorola, but I might be wrong about that.

I don't think so. I don't remember a connection, at least, but it'll be in here straight from the horse's mouth:

https://computerhistory.org/blog/transplanting-the-macs-cent...

I watched all of that a while ago and thought it was very interesting. Recommended.

Same! Thanks Jim for all your work on A/UX back in the day. You made a huge difference.

Reply on news.ycombinator.com