Hacker News Reader

Comment by yjftsjthsd-h | Hacker News Reader

> To start off the install, we begin with the “System setup and README” disk. We need to partition the disk, and then do something counter-intuitive: install System 6 on a Mac partition. This is because there’s a Mac application that kicks off the A/UX boot process: SASH; the A/UX standalone shell. This ‘pre-boot environment’ allows for launching an A/UX kernel and also some disk and recovery operations.

Funny how that rhythms with having a macOS install next to Asahi Linux. The more things change:)

Also, swapping through 26 floppies to install would have been... Something.

View on HN

> Also, swapping through 26 floppies to install would have been... Something.

Windows 95 was about that size, and Office was closer to 50?

At my very first job I remember installing stuff that way...ugh

I remember installing Office on Windows 3.1 and it was about 35 floppies at that time IIRC.

If I remember correctly, there's an interesting historical reason for this: a lot of the original functionality that we'd today consider "part of the OS" was actually in ROM on hardware in really old Macs. Mouse functionality, basic windowing, etc. This meant that to get A/UX running you first had to bootstrap into a light version of Mac OS and then boot into A/UX.

The Toolbox ROMs, right? I can see the utility of using that (I mean, beyond that you might need to use it to boot), but why couldn't A/UX call those APIs itself? I can easily see where bootstrapping through Mac OS would be easier, but I can't immediately see why it would be particularly necessary.

It's been a long time, but I'm pretty sure A/UX didn't use any of the toolbox roms and had it's own drivers (we had the source). A/UX booted from a MacOS partition because the Mac bootloader only understood booting MacOS (and it wasn't writeable with new boot code), so you booted to MacOS, then started SASH, which loaded Unix.

That's exactly the reason. NetBSD uses its own booter for the same purpose, for example.

Yeah, that was similar to Amiga and the Kickstart concept, initially on floppy, then as a separate ROM module. Going from AmigaOS 1.3 to 2.0 with the applicable Kickstart ROM gave you a whole new UI layer.

I miss the Amiga.

Yes. It was a pretty big chunk, too. Floppies were around 880k and base model RAM sizes were 512k to few MB. Having better part of a meg of libraries in ROM really helps fit more stuff in. And it was possible to load RAM patches or shadow things IIRC.

https://www.amigaforever.com/kb/15-127

RAM: and RAD: the RAM drive and the Recoverable RAM Drive - because putting things in memory that can survive a reboot isn't just for virii!

It is sort of the same for DOS. When you start digging around in the source you realize it is only really half an operating system and a surprising amount is done in the BIOS.

Yes, and it really was just a Disk Operating System.

Not really (I wrote the mouse/display drivers and kernel event queue driver for A/UX 1.x) - A/UX has it's own kernel ADB (and mouse/kbd on top of that) and display drivers, it will happily boot from hard drive, and throw up a non-mac terminal on a Mac display card without executing anything from the Mac ROMs.

The later MacOS running on A/UX ran in a single A/UX (system V) process

So what am I remembering that A/UX needed to boot from ROM?

I'm pretty sure the ROMs loaded a boot sector booter and jumped into that, it's been a long time but I'm pretty sure that's what we handed over to Apple

I guess that makes sense. Thanks for the answer and the trip down memory lane.

It's possible that later A/UXs required booting MacOS - but when we started we got 1/2 the initial prototype production run of Mac 2s (no cases, bugs in pals, probably no standard ROMs yet)

Our standard agreement for porting Unix to new hardware (we did something like 150 different boxes, not just macs) required a minimal ROM monitor that could download code from a serial port (we could provide source), we'd take it from there, write S/A code to store data on the harddrive, download a file system, start downloading kernels into ram debugging the disk and serial drivers until we got a shell prompt ... then moving development onto the box

Also, swapping through 26 floppies to install would have been... Something.

We installed it from a QIC tape when it wasn't delivered on a SCSI hard drive. Not sure if that option was generally available tho; we were doing kernel development.

Installed via QIC tape to Rodime 120Mb drive. Had fun formatting the Rodime; but got it working.

Learnt TCP/IP with this setup. Good Times(tm)

Yeah...using anything other than an Apple Approved(tm) scsi disk (or, really, anything 'not Apple') was an adventure. One of my least favorite Unixen, frankly. SVR2 based when most of the world on the AT&T side of Unix was SVR3 (and SVR4 was just around the corner), some BSD bits tossed into the blender, coated in a not-very-good MacOS emulation layer. Congrats to you on your perseverance; I recall the TCP/IP stack was...challenging.

A/UX was created so Apple could bid on government contracts that had a "must be Unix and/or POSIX" checkbox (common requirement at the time). IMO, they should have either gone all-in on a "MacOS apps running perfect on a Unix kernel" and picked a better kernel, or done the Atari thing and tossed the legacy OS and gone all in on a "Unix on Apple hardware". Instead it was 'meh' all around.

> Also, swapping through 26 floppies to install would have been... Something.

Perfectly normal.

You would need a few for MS-DOS (I think only MS-DOS 3.3 was the last that could fit into an high density one).

Followed by about 10 for Windows 3.x, similar numbers for Turbo Pascal, Turbo C++, Clipper, Word, Excel (replace with the competition),....

And before installing all of that, don't forget to as very first step, create backup floppies from all of them, and actually use the backups during the installations, while keeping the originals in a safe place, not that the drive could damage one of them.

> Also, swapping through 26 floppies to install would have been... Something.

I still have a legit copy of Word on 10 floppies.

It was bad. And when you'd copy so many floppies, typically one would fail and you'd only notice when installing. We weren't very advanced back then (at least I wasn't): no fancy an 11th "parity" disk that'll fix any other one that'd fail. At least not for me.

The data CD-ROM was a very welcome addition to the world back then.

I remember a trick (I think it was for Windows 95) where you could copy all the files to the hard drive first and then once they were all there, you could install directly from them.

A/UX 1.0 came on a pre-written 80MB disk, which indeed would have been a lot easier.

Oh, that's even better:) I assumed the mentioned tape was the preferred way to get it, but a no-op install is faster/easier yet!

80Mb Drives were _expensive_

I don't get the impression that a single aspect of A/UX was meant to be cheap, except perhaps in comparison to the other Real UNIX Workstation offerings.

Reply on news.ycombinator.com