I really look forward to reliable (modern) macOS emulation, a la wine. I miss some of my wee little Mac utilities / games now that I’m in Linux full-time. I’m rooting for the guys working on stuff like Darling[0], but it looks like they still have a long road ahead.
I took a look at their page, and they say that they "finally have basic experimental support for running simple graphical applications". However, their blogs/progress reports are 4 years old. Do you know how far along their support for graphical apps is? Running GUI apps is what's going to make this useful, and is what nobody else (AFAIK) was ever able to do.
I agree that GUI apps are the real use case here. Honestly, reversing and recreating the MacOS APIs is such a mammoth effort though, I can understand why progress would be slow. Wine faced a similar struggle, but also had an urgent need driving it: we needed Win apps to work to make Linux viable for most people. That’s less the case for Mac apps, so Darling is by necessity more of a passion project.
When I clicked link#2 above I got nostalgic. The URL had my mind conjuring up images of AOLs CDN, powered by a massive array of beige Quadra 650s with AOL discs in their CD-ROM drives.
The proper name is Mac OS 9, you are never going to convince people to call it by something else just to make a far more obscure operating system easier to find on Google.
If I search for “OS 9” without any quotes or anything I get the Wikipedia article for Mac OS 9. I think it’s fine. There’s enough niners out there keeping things alive.
I think that was what the poster was implying was the problem. MacOS 9 and OS-9 are not the same thing. OS-9 is a Real Time OS originally for 6800/68000 processors.
We have Haiku which is getting quite good and Amiga is still in production, their new computer that is currently in beta testing looks pretty good, they would possibly get me to part with the money if they released a laptop.
> Amiga is still in production, their new computer that is currently in beta testing looks pretty good
The Amiga "Tabor" or "A1222" is a very, very silly machine. Because they insist on PowerPC even though that's basically a long dead architecture, the best available CPU for it was the QorIQ P1022 a SOC designed to be inside say a mid-range Ethernet switch a decade ago.
In the event they actually ship a product in some sort of volume (say, dozens) before NXP discontinue the processors altogether this would be far worse than a PC or Mac you could have bought in 2015 when it was conceived, and yet also far more expensive, very silly.
It is very very silly of you to think they are trying to compete with modern general purpose desktops, they are attempting to fill a niche which Amiga helped create and has never quite been filled since their demise despite modern computers being so much "better."
I was referring too the X5000 which is PowerISA architecture. A1222 is planned to be around $500 which is not all that expensive, X5000 is on the expensive side. Considering my use of such a computer would be to run various audio software that was designed for much slower single core machines I suspect a dual 2ghz processor with XMOS coprocessor would be quite satisfactory. It would allow me to use some software which does not play well with emulators and would probably be just about ideal for my needs when it comes to a studio computer.
I suspect the A1222 project was scrapped due to shutdown fallout, chip shortages, etc but who knows, it may still happen.
The Amiga was one of several very popular late 1980s home computers. They all have this same rather vague nostalgic sentiment attached to them by fans, but it's true that none of them have been more preyed upon by crooks than the Amigans over the subsequent years.
Nothing is going to "fill that niche" because it's not actually a niche, just your brain playing tricks on you.
And the X5000 isn't "in production" in any meaningful sense. Somewhere there may be some boards on a shelf that are technically "new in box", but this is fairly old hardware from an even older design. That P5020 CPU is thirteen years old and whilst its industrial customers presumably used the second core, Amiga software does not.
Amiga "developers" periodically claim that they're working very hard on that, all the way back to 2013 you can find very earnest sounding men explaining that they've mostly got it working but it is being tested by a select group under NDA. They're bullshitting, telling Amigans you just need a few more weeks a bit more money means you don't have to admit that you spent every penny and there's nothing to show for it. Maybe a miracle will happen.
Amiga OS does use the second core even if the software does not, just like every modern system. Do you think it just piles every single process onto a single core?
I never said the x5000 was in production, I actually said it wasn't.
The closest thing to "Amiga Developers" would be Hyperion who took over the OS and has made licensing deals with 5 or 6 different hardware producers, they make no hardware themselves. Over the past decade or so at least a dozen different Amiga systems and motherboards have gone into production and sold well for what it is. Most of the hardware is sold through various online stores with a couple of those producing hardware also selling online, Hyperion only sells software; so who exactly are these evil "Amiga developers" you talk about? Or is it a conspiracy and they are all involved?
If you are going to make accusations you should identify the culprit, would you blame Microsoft and everyone who makes or sells computers which run Windows because HP ripped you off?
> Amiga OS does use the second core even if the software does not, just like every modern system. Do you think it just piles every single process onto a single core?
It's much worse than that. Only the first core is even "running". The Amiga OS isn't intended to handle true concurrency, which is why the 68k era operating system provides Forbid/Permit and Disable/Enable, which respectively forbid the OS from task switching and disable all interrupts. This made writing a 1980s home video game for the Amiga hardware much easier, but most application software took advantage of this too.
The PowerPC port of the system from the turn of the century actually doubles down on this nonsense, presumably because it made it easier to implement. As a result even "new" Amiga software, such as it is, still relies on these Forbid/Permit and Disable/Enable semantics, ensuring it can't be made to run efficiently on multi-core hardware. If there was widespread new application development you could imagine first shipping new SDKs which avoid these semantics, and latterly getting rid of the programs which use them but obviously for Amiga that's not practical at all and probably hasn't been for many years.
Over the 20+ years since I'm afraid things got a lot more complicated than "Hyperion took over the OS" although since Hyperion are one of several entities engaged in pointless legal wrangling about this topic perhaps if you believe their version of events that's what you'd get out of it. Trevor Dickinson ends up funding a lot of the remaining software development. sometimes via the A-EON Technology Ltd vehicle, other times personally.
There are a long parade of "culprits" if you want to use that term, whether Amiga Inc, Hyperion Entertainment, Eyetech Ltd, or specific individuals like the Friedens, Bill McEwen, Steven Solie, Ben Hermans, Alan Redhouse, Trevor, there are always people who'd like to exchange your money for their promises. Do they know the promises are bogus? I don't care, and if you ensure you don't give them money then nor should you.
I installed Haiku on an old Eee and yes it runs like dream and I haven't experienced single kernel panic yet. I'm thinking about installing it on something "bigger" than Eee as I was really surprised how usable and nice the OS is.
That's not normal behavior (these days, anyway!) A lot of users are able to go months without kernel panics, especially on certain kinds of older hardware.
I can't remember if I learned about BeOS on ZDTV or a magazine or where, but I do remember going in to Best Buy with the intention of buying it as a nascent teenager with a license. My heart broke a little that I couldn't just walk in a store and get it, but I still ordered it and learned a lot about dual booting my parents' Packard Bell thanks to Be. I loved how it was focusing on being a media operating system, with good graphics and audio processing capabilities, and I was a kid who loved music and the radio. If only it had survived.
There was a version of BeOS (iirc, it was the last version they released) that could be installed in a file on a Windows9x hard drive and booted to by using a boot floppy.
It really was an impressive and innovative way to get people to try a new OS, keeping in mind that most people couldn't (and still can't, to be fair) set up a more normal dual boot setup. I think they were just too close to shutting down by that point, so nothing ever really came of it.
Neal Stephenson's In the Beginning Was the Command Line speaks quite favorably of BeOS. He described the Windows world as being a kind of car dealership that sold clunkers, Linux as a place where anyone could walk in and buy a US Army M1 Tank, and BeOS as fully operational Batmobiles.
In those days Linux was an M1 that sort of but not quite ran on your local fuel and was 1/2" wider than your garage. The fuel problem would be sorted out just as soon as the engineers can get your city to admit what they added to the fuel but they won't even admit having knowledge of anything called "fuel." Eventually you get the new part that will make your M1 run on your local fuel but you need to rebuild the engine to install it, you don't really have any clue of how to do that but you have been talking to some vague personas in an alley for the past 3 nights and you are fairly certain you can do it. All seems to go well until you try and reinstall the engine, somehow it has gotten considerably larger than it had been and no longer fits in the engine compartment. So you rebuild it again, and again, and again, finally it works but now the glove compartment light is flashing error codes in Morse and the radio constantly drifts out tune. Repeat once or twice a year until about 2005 or so.
But you have your M1, you can not buy ammunition for its guns or drive it on any roads because the tracks will eat up the pavement but it is your M1.
I merely quoted -- or at least paraphrased -- the original essay itself:
«
They've been modified in such a way that they never, ever break down, are light and maneuverable enough to use on ordinary streets, and use no more fuel than a subcompact car. These tanks are being cranked out, on the spot, at a terrific pace, and a vast number of them are lined up along the edge of the road with keys in the ignition. Anyone who wants can simply climb into one and drive it away for free.
»
This essay is so important, so valuable, and so blasted insightful, that even though I read it more than 20 years ago, I can still remember key phrases, and search for it, find it, and quote from it in seconds.
It should be one of the seminal texts of the FOSS movement. And yet most people don't know it, or if they do, they don't get it and misquote it or misinterpret important points... such as here, claiming that people were trying to sell Linux up against BeOS and MacOS and Windows, when the real point was that although it was better than all the commercial competition in important ways, it was FREE and yet the people that developed it literally could not GIVE IT AWAY for nothing, up against inferior commercial OSes.
And they still can't.
Everyone has probably also forgotten by now that Sun bought StarOffice and tried to give it away, and nobody cared. So some forgotten Sun exec, who actually understood how businesspeople's minds work, made a discount version. It was $89.95 or something, alongside and as well as the free version.
The result was a modest explosion in business uptake.
A free office suite? Must be junk. But a bargain basement full-function office suite, for eighty bucks when MS Office costs a few hundreds? That's a steal! Great -- I will have that and push it out across my business!
It worked.
And now, Linux is on the desktop and it's a roaring success, selling hundreds of millions of units, and outselling Macs.
It's called ChromeOS, and when I point this out, the geeks of HN get very angry at me, and complain, and say that I'm moving the goalposts, that ChromeOS doesn't count, that it's
not really Linux.
Even one specific commenter I really respect and whose posts I find richly informative and learn from, flat out deny that it is Linux:
https://news.ycombinator.com/item?id=34971384
Ah well.
We all know Linux runs servers, routers, and most smartphones, but if it's on the desktop, and it costs money, then it's obviously not Linux. If it's a successful desktop that people like, it clearly can't really be Linux, right?
Shakes head, walks away.
Anyway, for better context, here's the whole passage from ITBWTCL:
«
With one exception, that is: Linux, which is right next door, and which is not a business at all. It's a bunch of RVs, yurts, tepees, and geodesic domes set up in a field and organized by consensus. The people who live there are making tanks. These are not old-fashioned, cast-iron Soviet tanks; these are more like the M1 tanks of the U.S. Army, made of space-age materials and jammed with sophisticated technology from one end to the other. But they are better than Army tanks. They've been modified in such a way that they never, ever break down, are light and maneuverable enough to use on ordinary streets, and use no more fuel than a subcompact car. These tanks are being cranked out, on the spot, at a terrific pace, and a vast number of them are lined up along the edge of the road with keys in the ignition. Anyone who wants can simply climb into one and drive it away for free.
Customers come to this crossroads in throngs, day and night. Ninety percent of them go straight to the biggest dealership and buy station wagons or off-road vehicles. They do not even look at the other dealerships.
Of the remaining ten percent, most go and buy a sleek Euro-sedan, pausing only to turn up their noses at the philistines going to buy the station wagons and ORVs. If they even notice the people on the opposite side of the road, selling the cheaper, technically superior vehicles, these customers deride them cranks and half-wits.
The Batmobile outlet sells a few vehicles to the occasional car nut who wants a second vehicle to go with his station wagon, but seems to accept, at least for now, that it's a fringe player.
The group giving away the free tanks only stays alive because it is staffed by volunteers, who are lined up at the edge of the street with bullhorns, trying to draw customers' attention to this incredible situation. A typical conversation goes something like this:
Hacker with bullhorn: "Save your money! Accept one of our free tanks! It is invulnerable, and can drive across rocks and swamps at ninety miles an hour while getting a hundred miles to the gallon!"
Prospective station wagon buyer: "I know what you say is true...but...er...I don't know how to maintain a tank!"
Bullhorn: "You don't know how to maintain a station wagon either!"
Buyer: "But this dealership has mechanics on staff. If something goes wrong with my station wagon, I can take a day off work, bring it here, and pay them to work on it while I sit in the waiting room for hours, listening to elevator music."
Bullhorn: "But if you accept one of our free tanks we will send volunteers to your house to fix it for free while you sleep!"
Buyer: "Stay away from my house, you freak!"
Bullhorn: "But..."
Buyer: "Can't you see that everyone is buying station wagons?"
»
Yeah, true, it is, but still... I'm probably within 20y of death from old age, if I make it -- I'm now the age my dad was when he died. Has your memory suffered worse than mine over that time?
Not trying to be nasty but I mean, come on, it was THE key point of the essay.
I think there are a lot of valuable insights one can take away from Stephenson's essay (it's a favourite of mine as well), and one person's key point may be another person's side discussion.
Stephenson did compare Linux to a tank, but he also compared it to the Milwaukee Hole Hawg drill, and that analogy (rather than the tank, which I'd forgotten about entirely) was more memorable for me personally.
His argument was that Linux was a very powerful tool for people who know what they're doing. The Hole Hawg has a motor that's so powerful, if the bit gets stuck, the body of the drill will spin in the other direction with enough force to spin the operator around (if they manage to hold on) or break their arm(s) if they lose their grip. People doing certain kinds of serious work want that level of power, but everyone else would most likely be happier with a regular power drill. Not only are they safer, they're a lot smaller and weigh a lot less.
Similarly, getting a tank with great gas mileage for free sounds pretty neat at first glance. On the other hand, if someone used to driving a sedan tries to use a tank as a drop-in replacement, the regrets are going to pile up pretty quickly. They might collapse some structures designed for regular vehicles just by driving over them, because tanks weigh so much more. They'll most likely tear up dirt and gravel driveways. They probably won't be able to park it in the places they'd park their car, because the tank is so much bigger. Getting in and out of the vehicle is more of an adventure.
There are definitely folks who would love their own tank - I know in Australia there are people that converted surplus WWII tanks into farming equipment - but most non-specialists would be unhappy with one, just like they'd have been unhappy trying to use Linux as a desktop OS in the era when In the Beginning was the Command Line was written, regardless of the price.
BeOS felt like it was going to become a serious modern alternative OS for the PC focusing heavily on multimedia, similar to MacOS (Be was founded by ex Apple people so it makes sense.) Unfortunately it never happened. Thy pivoted to internet appliances, Palm bought them out, the end. I still have my R4 and R5 media somewhere.
It was seriously considered by Apple as the successor of the classical MacOS but they got too greedy with the price and Apple went with a NextStep based OS.
In all honesty BeOS had no leverage because they were so committed to build the successor of MacOS that they had no alternative plan
Be did get too greedy with the asking price, you're right.
But that was not the only thing.
Be had no particularly special story about dev tools; NeXT had the world's best, plus a charismatic, visionary CEO.
Be should have gone and done a deal with Acorn, who could have done them a silent-running, passively-cooled SMP multi-ARM workstation in 1998 for less than the cost of a mid-range Mac.
Acorn had cheap SMP hardware ready to go. Be had the state-of-the-art SMP workstation OS before Linux 2.2 made it able to scale usefully.
If only, I miss BeOS. ...
Though I think it would be the same chicken/egg APPS problem with the BeBox: Unless you could run existing Risc OS apps under BeOS, I doubt Acorn users would want to switch, and porting existing apps through a Carbon-like API subset would probably be hard, because I gather a lot of apps use assembly and hacks to the OS..
ps
I just started trying out Risc OS after reading a couple of your articles on it (modernizing (or not) and the gem of an os (or not)).
I've got it on my pi 400, and use it for writing without distraction (no wifi!), and there is no latency while typing. The ADJUST button thing is also pretty neat.
Realistically I don t think Acorn ever really and a hope of becoming a mass market success because their failure wasn’t technical. BeOS was a far better operating system than RiscOS but at the end of the day agglomeration effects are far more important than technical prowess. Maybe they could have captured some of Sun or SGI’s market share but who remembers them either?
(I say this as someone who grew up with first and Electron then an A3000.)
I reckon Acorn had 2 possible escape routes, but they didn't seem to notice either.
[1] Make the thinnest lightest longest-battery-life laptops around, able to do all the stuff the DOS/Win9x phat boyz did but with no fan, no extra cooling, and do it for a day at a time.
RISC OS would have been fine for that... but it was just a different dead end road.
[2] The Phoebe (RISC PC 2) was SMP-capable, had richer media support and PCI and so on, but RISC OS wasn't SMP ready and wasn't going to be.
SMP was already manifestly The Future. So, commit to it. Current SOTA OSes in ~1998 that could do SMP were either:
• Expensive proprietary UNIX workstations -- so undercut them
• Windows NT 4 -- not really finished yet, serious weaknesses (e.g. no USB), even if it was the first version with a credible UI.
• Linux was not there yet.
There was an opening, but the plan to build their own OS -- again -- was insane. They didn't know it but they needed a partner with a lightweight SMP media OS. Be had a lightweight SMP media OS but needed more competitive hardware.
But these are just daydreams.
If the bottom line is just maximising revenue -- to hell with changing the world, just sell more sugar water (to paraphrase Jobs to Sculley) -- then they did the right thing. Sell lotsa chips, dump the workstations and the OS.
Be didn't have anything. Developers, developers, developers, right?
Late 90s I was running OS X Server, Be, and NeXT. NeXT was better than OS X Server, and Be was far behind. Looked pretty, but that was about it. Funny thing is I live in a browser and text editor now, and the Be install I ran in 99 would just about do it!
NeXT was just more mature to me at the time. OSXS was trying to do too much from all directions, while I was of the belief that System 8.6.1 was pretty perfect already.
Agree on Be, and I've tried Hauki off and on. It really seems, as I default to using my phone for nearly everything anyway, that a system like that really is ideal at this point. Like a lot of people my day is spent a jump server away, so I haven't cared what I'm using 99 percent of the time.
I mean, yes, NeXTstep was much more mature. No argument. But OS X Server was aimed at a different job, and a different job to classic MacOS as well.
People did use NeXTstep as a server -- the first ever web server, notably -- but it wasn't a great one and wasn't its intended role.
Classic MacOS was usable as a server too. It seems to be largely forgotten now but the US Army officially moved all its web servers to classic MacOS at one point, because not having any form of shell at all made it largely unhackable. You can't get shell access when there is no shell to get to, and no clever x86 code exploit will help you when none of the servers have an x86 chip in them.
It wasn't a very good server, though, and quietly I think they changed to something else a bit more conventional not long afterwards.
Compared to Classic, I bet OS X Server was a great server OS. ;-)
I think the real purpose of OSXS was to answer the question:
"OK, we've done a ton of work porting NeXTstep to PowerMac and grafting a kinda sorta MacOS-like desktop on the thing, and it's cost us a sh1t-tonne of money and person-hours of work with not much to show for it so far. Is there any way we can start to make some money off this thing before it's ready to ship as a desktop OS?"
The release of BeOS in 1995 pretty much marked the end of desktop operating systems diversity for nearly two decades. Windows 95 had already been released and quickly cornered the market. Although we perhaps have fewer options today than in the 80s and early 90s things are pretty healthy. No one operating system is dominant in the same way. Alongside Windows, ChromeOS and macOS both have significant shares on desktop and for consumers iOS and Android are just as important.
There's also Mac-on-Linux (MOL), which I've been using from ~1999 to run the original OS under Linux on my Mac hardware until I dropped Mac OS completely a few years later. It does require a PPC host. I remember back then MOL was preferable to SheepShaver, but I don't remember why anymore.
> I remember back then MOL was preferable to SheepShaver, but I don't remember why anymore.
Speculation: Running on native hardware probably give you better performance, and probably meant that you didn't have to provide a ROM file because it could just read the chip that's actually installed in the machine.
Sheepshaver runs natively on supported Macs under BeOS and you likewise don’t need a ROM file as it uses the one built in to your Mac. The real reason is probably it doesn’t support MacOS past 8.5 really. Not well.
I had MacOS 9, so that, performance, may have been part of it. Seems MOL emulated an MMU. Although, IIRC I ended up disabling virtual memory / swapping inside MacOS, and just give the emulated Mac more RAM, because doing the swapping purely on the Linux side was more performant. I liked the idea that Linux was simply more powerful, although it might have just been the interaction between the two caches. :)
I remember I was also very pleased to get copy-paste working between the systems, but I may have had to install something for this.
SheepShaver does not run MacOS X, just the classic, while MOL runs both. Been using MOL about at the same time to check out OS X, which led me buying a Pismo, my first and last Apple machine.
It’s probably worth mentioning that you can also use QEMU, which is a bit more involved but can use a bigger disk and also run OS X, and Mac OS 9.2.2 (as it supports an MMU, SheepShaver doesn’t)
Is QEMU the best option to run older OS X (in particular Snow Leopard, and yes, a legal retail purchased physical copy) on a modern machine? I currently keep a rather bulky cheese grater to run a few programs from time to time.
That's the option with the least headache. I believe there are other simulators too but they are much more specialized.
With some hacking you can run all MacOS/MacOS X versions on x86, x64 and PPC from 8.6 to 13 if you have a x64 machine. If you have an Apple Silicon one it gets too slow to simulate around 10.10 (on x64 you can use hvf virtualization). It might be a bit less nice than tailored emulators like SheepShaver but having the same UI for all these OSes (plus other OSes) beats everything in my opinion.
Unfortunately Apple Silicon machines cannot be simulated without some hidden knowledge only some secretive companies have. Hopefully the knowledge will become public some years from now.
My memory is that it only had to be run on Apple hardware, but perhaps I should go back and check. Or just not tell anybody, and hope the Apple Police don't get me — it's not like I'm letting something that old out to play on the internet (though I do sometimes miss iCab).
Unfortunately qemu isn't as 'integrated' as Sheepshaver -- theres the mouse grab problem, also no disk sharing with the host (the read[/write] fat32 disk don't seem to work, at least under 9.2.2)
I had a a project for #marchintosh of making myself a nice powerPC emulation stack to recompile and open source quite a few free/shareware I did back then, but I more or less gave up as I couldn't find something 'comfortable' to work in. I guess I'll have to find, dust up and try to hook that old G4 Mac Mini that I'm sure is in a drawer, somewhere :-)
You can, but not under BeOS. Under BeOS Mac OS is run almost like Classic was on MacOS X. The OS is more or less virtualised as a user space app, and the CPU is not emulated in any way. That’s the main reason that BeOS only supports PowerPC. Indeed, if you run it on a Mac under BeOS it doesn’t even need a toolbox ROM
SheepShaver was the successor to the Amiga's ShapeShifter Mac emulator (also written by Christian Bauer). I had fun running Mac OS and games on my Amiga. It also had the interesting property that my Amiga had a 68060 processor, which no Mac had, meaning that it was faster at running 68k apps than any official hardware was.
I used Sheepshaver to emulate system non-PowerPC games + System 7.5.3 when making Mac game nostalgia videos on my YouTube many years back. I found it to be far more performant than Basilisk for anything that wanted to run with Quickdraw. It also seemed to have less clipping audio.
I should see how it runs all these years later on a Linux box.
But this requires a PowerPC processor under BeOS so you can’t use this with Haiku at the moment. The Linux version looks like it emulates the CPU so it can probably be made to work I guess?
No, not really. It isn’t just PowerPC, it is a very specific and picky Mac, has PCI, is old world, has a 603 or 604 processor when shipped originally. If you have one, you then need a copy of BeOS R4.5 or R5, and getting that installed can actually be non trivial (as ActionRetro proved on YouTube - it was my custom pre built image that he finally got working.)
I have 2 BeOS PowerPC boxes, so yep - know all the hassles first hand.
If you do too - awesome, you should come find me on Haiku forums and we can contribute to the BeOS PowerPC preservation effort I started last year.
By the time my family got a home computer it was a 120 Mhz Pentium system, but I spent basically all of elementary, middle and high school using m68k and PPC macs at school. During the pandemic i picked up two m68k Macs and 2 PPC Macs to satisfy my nostalgia. I enjoy playing with old tech, but theres nothing wrong if you don't! Different strokes and all that.
Ah, please don't get me wrong: I love playing with old tech too. The first computer I used was some 386 or 486 machine running Windows 3.11, it even had a turbo button and a 5.25" floppy drive. Really wish we still had that thing.
It's just my path never crossed PowerPC's, so I don't have the nostalgia for them that I do for classic x86 hardware. Regardless I'm sure they are an important piece of computing history.
Near as I can tell the POWER9 chips (at least the IBM ones) only have backwards compatibility through POWER7 from 2010. The last PPC chip Apple used, the G5, implemented the POWER4 ISA, so that's pretty far back.
Nope. It really never even supported G3 officially. It will boot on the BeBox natively an a Mac with a boot loader. Without modifying the kernel, I doubt it will boot on much else.
There is also currently no way to even virtualize BeOS for PowerPC. I wish there was. The closest I have seen is a MESS that could boot and show the boot splash screen (this is a 3d Be logo on PowerPC, the one you see on Intel was only ever on Intel.) The dingusppc have BeOS as a possible target for their Mac emulation, but it is not there yet. I'm pretty sure QEMU can't boot BeOS because BeOS requires an old world ROM and it requires a PCI Mac with a 603 or 604 processor - so a 4400, 7300, 8500/8600, 9500/9600 and a few others (including clones) work - but nothing NUBUS and nothing post the first motherboard revision of the 8600/9600. The later x600's have a different motherboard and do not work. Nothing that shipped with a G3 and nothing with new world ROM.
Instead of emulation like this (which is really cool, BTW), I'd love to see something like GNUStep or Etoile growing into a full implementation of all macOS APIs so that I could just build a native Mac application against it and have it run on Linux.
I think that it's often underestimated just how much of what made/makes OS X/macOS good is actually in AppKit and the various other frameworks it's used with. Parts of it are a little aged compared to something that's received more attention in the past decade (like UIKit), but it still does a LOT of heavy lifting and makes developing a polished, capable desktop app much less of a headache.
In writing apps with other desktop UI frameworks, it's become evident just how much AppKit does well that many frameworks barely do or don't offer at all. Sometimes it's even really basic things, like how WinUI somehow has no first-party tableview despite that being one of the most fundamental desktop widgets.
Indeed. It’d need to be extended to cover the various other subsystems and probably offer a less idiosyncratic look and feel.
Not that I dislike the NeXT visuals - I’d totally have a NeXT as my daily driver (with X, of course), but it became dated the moment Windows 95 copied it.
In any case, anything that can be documented can be implemented in a clean room.
The use of "macOS" capitalization somehow makes it appear as if it's about the current Apple OS, when in fact this is about classic MacOS.
I really look forward to reliable (modern) macOS emulation, a la wine. I miss some of my wee little Mac utilities / games now that I’m in Linux full-time. I’m rooting for the guys working on stuff like Darling[0], but it looks like they still have a long road ahead.
[0] https://darlinghq.org/
I took a look at their page, and they say that they "finally have basic experimental support for running simple graphical applications". However, their blogs/progress reports are 4 years old. Do you know how far along their support for graphical apps is? Running GUI apps is what's going to make this useful, and is what nobody else (AFAIK) was ever able to do.
Their GitHub is active, and they have a Discord. I don’t do Discord, but if you do, you could probably follow their progress on there.
As of 2020, graphical app support appeared to be basic but somewhat functional: https://github.com/darlinghq/darling/issues/657
I agree that GUI apps are the real use case here. Honestly, reversing and recreating the MacOS APIs is such a mammoth effort though, I can understand why progress would be slow. Wine faced a similar struggle, but also had an urgent need driving it: we needed Win apps to work to make Linux viable for most people. That’s less the case for Mac apps, so Darling is by necessity more of a passion project.
Mac OS, to be pendantic about styling :)
Nope.
No space.
MacOS == System 1 to MacOS 9.
Mac OS == OS X == 10.0 to 10.12.
macOS == 10.13 to 13.
> MacOS == System 1 to MacOS 9.
I can't speak to previous releases, but Mac OS 9 definitely used a space. (Before looking it up I would've bet on "no space" too, BTW.)
https://imagej.nih.gov/ij/docs/install/images/about.gif
I believe the space was added when they formally renamed the operating system Mac OS in 1996 (with System 7, specifically 7.6).
System 7.5.1 and later used the name Mac OS in its boot screen [1] but the boxes still said “Macintosh System 7.5” [2] up until the release of 7.6.
[1] https://c1.staticflickr.com/1/12/69888558_ee83d000f9.jpg
[2] https://o.aolcdn.com/images/dims?image_uri=https:%2F%2Fs3.am...
When I clicked link#2 above I got nostalgic. The URL had my mind conjuring up images of AOLs CDN, powered by a massive array of beige Quadra 650s with AOL discs in their CD-ROM drives.
I do miss System 7.
Mac OS always had a space. Before it was called Mac OS, it was called the System Software.
> Mac OS always had a space.
[[Citation needed]]]
> Before it was called Mac OS, it was called the System Software.
Yes, true.
Actually there was a period when they dropped Mac altogether and called it "OS X" (Mountain Lion era)
My friends and I used to pronounce this "Oh Ess Ex", which I now know to be incorrect[0]
It's a hard habit to drop.
[0]: https://apple.stackexchange.com/questions/2549/whats-the-pro...
Good point!
> MacOS == System 1 to MacOS 9.
I seem to recall that versions prior to 7.5 were "Macintosh System Software".
True. I probably would have been better off saying "MacOS == 'Classic'."
Yes, if I remember correctly this can only run Mac OS 9. Strange that this is not mentioned (?) on the github page.
It can run up to OS 9, not just OS 9. It'll run any classic Mac OS that supports PPC, so I think that takes you back to the 7.5 line.
Please don't call it "OS 9" on the Internet. Googling for OS-9 resources is difficult enough as it is.
<https://en.wikipedia.org/wiki/OS-9#Name_conflicts_and_court_...>
The proper name is Mac OS 9, you are never going to convince people to call it by something else just to make a far more obscure operating system easier to find on Google.
If I search for “OS 9” without any quotes or anything I get the Wikipedia article for Mac OS 9. I think it’s fine. There’s enough niners out there keeping things alive.
I think that was what the poster was implying was the problem. MacOS 9 and OS-9 are not the same thing. OS-9 is a Real Time OS originally for 6800/68000 processors.
It's not really an issue. Searching for os 9 gives one OS's results. Searching for mac os 9 gives the other OS's results. Super easy.
The issue is that searching for "os 9" also gives results for Mac OS 9, making searching for OS-9 results harder.
Thanks for the explanation. In that case, the following search works:
-mac os 9
[dead]
7.1.2 was the first version that ran on PPC
>Strange that this is not mentioned (?) on the github page
"run classic PowerPC Mac OS applications"
"Runs MacOS 7.5.2 thru 9.0.4. MacOS X as a guest is not supported."
I see it now as well, strange that I completely missed that. Thanks.
The reference to BeOS brings me good memories. It was a time then the world of desktop operating Systems was much bigger than today.
BeOS, Syllable, SkyOS, Atheos, ReactOS, QNX etc.
At its time I liked BeOS phylosophy more than that of Linux and Windows, it's a shame it didn't take off.
We have Haiku which is getting quite good and Amiga is still in production, their new computer that is currently in beta testing looks pretty good, they would possibly get me to part with the money if they released a laptop.
> Amiga is still in production, their new computer that is currently in beta testing looks pretty good
The Amiga "Tabor" or "A1222" is a very, very silly machine. Because they insist on PowerPC even though that's basically a long dead architecture, the best available CPU for it was the QorIQ P1022 a SOC designed to be inside say a mid-range Ethernet switch a decade ago.
In the event they actually ship a product in some sort of volume (say, dozens) before NXP discontinue the processors altogether this would be far worse than a PC or Mac you could have bought in 2015 when it was conceived, and yet also far more expensive, very silly.
It is very very silly of you to think they are trying to compete with modern general purpose desktops, they are attempting to fill a niche which Amiga helped create and has never quite been filled since their demise despite modern computers being so much "better."
I was referring too the X5000 which is PowerISA architecture. A1222 is planned to be around $500 which is not all that expensive, X5000 is on the expensive side. Considering my use of such a computer would be to run various audio software that was designed for much slower single core machines I suspect a dual 2ghz processor with XMOS coprocessor would be quite satisfactory. It would allow me to use some software which does not play well with emulators and would probably be just about ideal for my needs when it comes to a studio computer.
I suspect the A1222 project was scrapped due to shutdown fallout, chip shortages, etc but who knows, it may still happen.
The Amiga was one of several very popular late 1980s home computers. They all have this same rather vague nostalgic sentiment attached to them by fans, but it's true that none of them have been more preyed upon by crooks than the Amigans over the subsequent years.
Nothing is going to "fill that niche" because it's not actually a niche, just your brain playing tricks on you.
And the X5000 isn't "in production" in any meaningful sense. Somewhere there may be some boards on a shelf that are technically "new in box", but this is fairly old hardware from an even older design. That P5020 CPU is thirteen years old and whilst its industrial customers presumably used the second core, Amiga software does not.
Amiga "developers" periodically claim that they're working very hard on that, all the way back to 2013 you can find very earnest sounding men explaining that they've mostly got it working but it is being tested by a select group under NDA. They're bullshitting, telling Amigans you just need a few more weeks a bit more money means you don't have to admit that you spent every penny and there's nothing to show for it. Maybe a miracle will happen.
Amiga OS does use the second core even if the software does not, just like every modern system. Do you think it just piles every single process onto a single core?
I never said the x5000 was in production, I actually said it wasn't.
The closest thing to "Amiga Developers" would be Hyperion who took over the OS and has made licensing deals with 5 or 6 different hardware producers, they make no hardware themselves. Over the past decade or so at least a dozen different Amiga systems and motherboards have gone into production and sold well for what it is. Most of the hardware is sold through various online stores with a couple of those producing hardware also selling online, Hyperion only sells software; so who exactly are these evil "Amiga developers" you talk about? Or is it a conspiracy and they are all involved?
If you are going to make accusations you should identify the culprit, would you blame Microsoft and everyone who makes or sells computers which run Windows because HP ripped you off?
> Amiga OS does use the second core even if the software does not, just like every modern system. Do you think it just piles every single process onto a single core?
It's much worse than that. Only the first core is even "running". The Amiga OS isn't intended to handle true concurrency, which is why the 68k era operating system provides Forbid/Permit and Disable/Enable, which respectively forbid the OS from task switching and disable all interrupts. This made writing a 1980s home video game for the Amiga hardware much easier, but most application software took advantage of this too.
The PowerPC port of the system from the turn of the century actually doubles down on this nonsense, presumably because it made it easier to implement. As a result even "new" Amiga software, such as it is, still relies on these Forbid/Permit and Disable/Enable semantics, ensuring it can't be made to run efficiently on multi-core hardware. If there was widespread new application development you could imagine first shipping new SDKs which avoid these semantics, and latterly getting rid of the programs which use them but obviously for Amiga that's not practical at all and probably hasn't been for many years.
Over the 20+ years since I'm afraid things got a lot more complicated than "Hyperion took over the OS" although since Hyperion are one of several entities engaged in pointless legal wrangling about this topic perhaps if you believe their version of events that's what you'd get out of it. Trevor Dickinson ends up funding a lot of the remaining software development. sometimes via the A-EON Technology Ltd vehicle, other times personally.
There are a long parade of "culprits" if you want to use that term, whether Amiga Inc, Hyperion Entertainment, Eyetech Ltd, or specific individuals like the Friedens, Bill McEwen, Steven Solie, Ben Hermans, Alan Redhouse, Trevor, there are always people who'd like to exchange your money for their promises. Do they know the promises are bogus? I don't care, and if you ensure you don't give them money then nor should you.
I installed Haiku on an old netbook recently.
It ran like a dream, except for kernel panicking every 5 minutes.
I installed Haiku on an old Eee and yes it runs like dream and I haven't experienced single kernel panic yet. I'm thinking about installing it on something "bigger" than Eee as I was really surprised how usable and nice the OS is.
That's not normal behavior (these days, anyway!) A lot of users are able to go months without kernel panics, especially on certain kinds of older hardware.
Did you report any of these panics?
I can't remember if I learned about BeOS on ZDTV or a magazine or where, but I do remember going in to Best Buy with the intention of buying it as a nascent teenager with a license. My heart broke a little that I couldn't just walk in a store and get it, but I still ordered it and learned a lot about dual booting my parents' Packard Bell thanks to Be. I loved how it was focusing on being a media operating system, with good graphics and audio processing capabilities, and I was a kid who loved music and the radio. If only it had survived.
There was a version of BeOS (iirc, it was the last version they released) that could be installed in a file on a Windows9x hard drive and booted to by using a boot floppy.
It really was an impressive and innovative way to get people to try a new OS, keeping in mind that most people couldn't (and still can't, to be fair) set up a more normal dual boot setup. I think they were just too close to shutting down by that point, so nothing ever really came of it.
Neal Stephenson's In the Beginning Was the Command Line speaks quite favorably of BeOS. He described the Windows world as being a kind of car dealership that sold clunkers, Linux as a place where anyone could walk in and buy a US Army M1 Tank, and BeOS as fully operational Batmobiles.
https://web.stanford.edu/class/cs81n/command.txt
You missed the point.
You could buy a Ferrari or a Ford, but an M1 Tank that ran for 1000 miles on a cup of fuel was for free. Keys in, sitting there. Get in, drive off.
In those days Linux was an M1 that sort of but not quite ran on your local fuel and was 1/2" wider than your garage. The fuel problem would be sorted out just as soon as the engineers can get your city to admit what they added to the fuel but they won't even admit having knowledge of anything called "fuel." Eventually you get the new part that will make your M1 run on your local fuel but you need to rebuild the engine to install it, you don't really have any clue of how to do that but you have been talking to some vague personas in an alley for the past 3 nights and you are fairly certain you can do it. All seems to go well until you try and reinstall the engine, somehow it has gotten considerably larger than it had been and no longer fits in the engine compartment. So you rebuild it again, and again, and again, finally it works but now the glove compartment light is flashing error codes in Morse and the radio constantly drifts out tune. Repeat once or twice a year until about 2005 or so.
But you have your M1, you can not buy ammunition for its guns or drive it on any roads because the tracks will eat up the pavement but it is your M1.
On the whole tanks don't have keys. Similar reason to airliners.
:-D Fair point.
I merely quoted -- or at least paraphrased -- the original essay itself:
« They've been modified in such a way that they never, ever break down, are light and maneuverable enough to use on ordinary streets, and use no more fuel than a subcompact car. These tanks are being cranked out, on the spot, at a terrific pace, and a vast number of them are lined up along the edge of the road with keys in the ignition. Anyone who wants can simply climb into one and drive it away for free. »
https://web.stanford.edu/class/cs81n/command.txt
This essay is so important, so valuable, and so blasted insightful, that even though I read it more than 20 years ago, I can still remember key phrases, and search for it, find it, and quote from it in seconds.
It should be one of the seminal texts of the FOSS movement. And yet most people don't know it, or if they do, they don't get it and misquote it or misinterpret important points... such as here, claiming that people were trying to sell Linux up against BeOS and MacOS and Windows, when the real point was that although it was better than all the commercial competition in important ways, it was FREE and yet the people that developed it literally could not GIVE IT AWAY for nothing, up against inferior commercial OSes.
And they still can't.
Everyone has probably also forgotten by now that Sun bought StarOffice and tried to give it away, and nobody cared. So some forgotten Sun exec, who actually understood how businesspeople's minds work, made a discount version. It was $89.95 or something, alongside and as well as the free version.
The result was a modest explosion in business uptake.
A free office suite? Must be junk. But a bargain basement full-function office suite, for eighty bucks when MS Office costs a few hundreds? That's a steal! Great -- I will have that and push it out across my business!
It worked.
And now, Linux is on the desktop and it's a roaring success, selling hundreds of millions of units, and outselling Macs.
It's called ChromeOS, and when I point this out, the geeks of HN get very angry at me, and complain, and say that I'm moving the goalposts, that ChromeOS doesn't count, that it's not really Linux.
They accuse me of "bad faith": https://news.ycombinator.com/item?id=34925697
They call my argument "BS": https://news.ycombinator.com/item?id=34928295
Even one specific commenter I really respect and whose posts I find richly informative and learn from, flat out deny that it is Linux: https://news.ycombinator.com/item?id=34971384
Ah well.
We all know Linux runs servers, routers, and most smartphones, but if it's on the desktop, and it costs money, then it's obviously not Linux. If it's a successful desktop that people like, it clearly can't really be Linux, right?
Shakes head, walks away.
Anyway, for better context, here's the whole passage from ITBWTCL:
« With one exception, that is: Linux, which is right next door, and which is not a business at all. It's a bunch of RVs, yurts, tepees, and geodesic domes set up in a field and organized by consensus. The people who live there are making tanks. These are not old-fashioned, cast-iron Soviet tanks; these are more like the M1 tanks of the U.S. Army, made of space-age materials and jammed with sophisticated technology from one end to the other. But they are better than Army tanks. They've been modified in such a way that they never, ever break down, are light and maneuverable enough to use on ordinary streets, and use no more fuel than a subcompact car. These tanks are being cranked out, on the spot, at a terrific pace, and a vast number of them are lined up along the edge of the road with keys in the ignition. Anyone who wants can simply climb into one and drive it away for free.
Customers come to this crossroads in throngs, day and night. Ninety percent of them go straight to the biggest dealership and buy station wagons or off-road vehicles. They do not even look at the other dealerships.
Of the remaining ten percent, most go and buy a sleek Euro-sedan, pausing only to turn up their noses at the philistines going to buy the station wagons and ORVs. If they even notice the people on the opposite side of the road, selling the cheaper, technically superior vehicles, these customers deride them cranks and half-wits.
The Batmobile outlet sells a few vehicles to the occasional car nut who wants a second vehicle to go with his station wagon, but seems to accept, at least for now, that it's a fringe player.
The group giving away the free tanks only stays alive because it is staffed by volunteers, who are lined up at the edge of the street with bullhorns, trying to draw customers' attention to this incredible situation. A typical conversation goes something like this:
Hacker with bullhorn: "Save your money! Accept one of our free tanks! It is invulnerable, and can drive across rocks and swamps at ninety miles an hour while getting a hundred miles to the gallon!"
Prospective station wagon buyer: "I know what you say is true...but...er...I don't know how to maintain a tank!"
Bullhorn: "You don't know how to maintain a station wagon either!"
Buyer: "But this dealership has mechanics on staff. If something goes wrong with my station wagon, I can take a day off work, bring it here, and pay them to work on it while I sit in the waiting room for hours, listening to elevator music."
Bullhorn: "But if you accept one of our free tanks we will send volunteers to your house to fix it for free while you sleep!"
Buyer: "Stay away from my house, you freak!"
Bullhorn: "But..."
Buyer: "Can't you see that everyone is buying station wagons?" »
yeah sorry it's been twenty years since this was published
Yeah, true, it is, but still... I'm probably within 20y of death from old age, if I make it -- I'm now the age my dad was when he died. Has your memory suffered worse than mine over that time?
Not trying to be nasty but I mean, come on, it was THE key point of the essay.
I think there are a lot of valuable insights one can take away from Stephenson's essay (it's a favourite of mine as well), and one person's key point may be another person's side discussion.
Stephenson did compare Linux to a tank, but he also compared it to the Milwaukee Hole Hawg drill, and that analogy (rather than the tank, which I'd forgotten about entirely) was more memorable for me personally.
His argument was that Linux was a very powerful tool for people who know what they're doing. The Hole Hawg has a motor that's so powerful, if the bit gets stuck, the body of the drill will spin in the other direction with enough force to spin the operator around (if they manage to hold on) or break their arm(s) if they lose their grip. People doing certain kinds of serious work want that level of power, but everyone else would most likely be happier with a regular power drill. Not only are they safer, they're a lot smaller and weigh a lot less.
Similarly, getting a tank with great gas mileage for free sounds pretty neat at first glance. On the other hand, if someone used to driving a sedan tries to use a tank as a drop-in replacement, the regrets are going to pile up pretty quickly. They might collapse some structures designed for regular vehicles just by driving over them, because tanks weigh so much more. They'll most likely tear up dirt and gravel driveways. They probably won't be able to park it in the places they'd park their car, because the tank is so much bigger. Getting in and out of the vehicle is more of an adventure.
There are definitely folks who would love their own tank - I know in Australia there are people that converted surplus WWII tanks into farming equipment - but most non-specialists would be unhappy with one, just like they'd have been unhappy trying to use Linux as a desktop OS in the era when In the Beginning was the Command Line was written, regardless of the price.
Yeah, that's a good, fair and well-reasoned analysis. Can't argue with that. :-)
BeOS felt like it was going to become a serious modern alternative OS for the PC focusing heavily on multimedia, similar to MacOS (Be was founded by ex Apple people so it makes sense.) Unfortunately it never happened. Thy pivoted to internet appliances, Palm bought them out, the end. I still have my R4 and R5 media somewhere.
It was seriously considered by Apple as the successor of the classical MacOS but they got too greedy with the price and Apple went with a NextStep based OS.
In all honesty BeOS had no leverage because they were so committed to build the successor of MacOS that they had no alternative plan
Be did get too greedy with the asking price, you're right.
But that was not the only thing.
Be had no particularly special story about dev tools; NeXT had the world's best, plus a charismatic, visionary CEO.
Be should have gone and done a deal with Acorn, who could have done them a silent-running, passively-cooled SMP multi-ARM workstation in 1998 for less than the cost of a mid-range Mac.
Acorn had cheap SMP hardware ready to go. Be had the state-of-the-art SMP workstation OS before Linux 2.2 made it able to scale usefully.
If only, I miss BeOS. ... Though I think it would be the same chicken/egg APPS problem with the BeBox: Unless you could run existing Risc OS apps under BeOS, I doubt Acorn users would want to switch, and porting existing apps through a Carbon-like API subset would probably be hard, because I gather a lot of apps use assembly and hacks to the OS..
ps I just started trying out Risc OS after reading a couple of your articles on it (modernizing (or not) and the gem of an os (or not)).
I've got it on my pi 400, and use it for writing without distraction (no wifi!), and there is no latency while typing. The ADJUST button thing is also pretty neat.
:-)
Yeah, you're right, it's just an idle daydream fantasy of mine.
RISC OS was and is pretty small and simple. I reckon doing to it what Apple did with Classic -- a version running in a native VM -- was doable.
Realistically I don t think Acorn ever really and a hope of becoming a mass market success because their failure wasn’t technical. BeOS was a far better operating system than RiscOS but at the end of the day agglomeration effects are far more important than technical prowess. Maybe they could have captured some of Sun or SGI’s market share but who remembers them either?
(I say this as someone who grew up with first and Electron then an A3000.)
This is a good point, yes.
I reckon Acorn had 2 possible escape routes, but they didn't seem to notice either.
[1] Make the thinnest lightest longest-battery-life laptops around, able to do all the stuff the DOS/Win9x phat boyz did but with no fan, no extra cooling, and do it for a day at a time.
RISC OS would have been fine for that... but it was just a different dead end road.
[2] The Phoebe (RISC PC 2) was SMP-capable, had richer media support and PCI and so on, but RISC OS wasn't SMP ready and wasn't going to be.
SMP was already manifestly The Future. So, commit to it. Current SOTA OSes in ~1998 that could do SMP were either:
• Expensive proprietary UNIX workstations -- so undercut them
• Windows NT 4 -- not really finished yet, serious weaknesses (e.g. no USB), even if it was the first version with a credible UI.
• Linux was not there yet.
There was an opening, but the plan to build their own OS -- again -- was insane. They didn't know it but they needed a partner with a lightweight SMP media OS. Be had a lightweight SMP media OS but needed more competitive hardware.
But these are just daydreams.
If the bottom line is just maximising revenue -- to hell with changing the world, just sell more sugar water (to paraphrase Jobs to Sculley) -- then they did the right thing. Sell lotsa chips, dump the workstations and the OS.
Be didn't have anything. Developers, developers, developers, right?
Late 90s I was running OS X Server, Be, and NeXT. NeXT was better than OS X Server, and Be was far behind. Looked pretty, but that was about it. Funny thing is I live in a browser and text editor now, and the Be install I ran in 99 would just about do it!
> Be didn't have anything. Developers, developers, developers, right?
Well, quite, yes.
> Late 90s I was running OS X Server, Be, and NeXT.
Early adopter, huh? OSXS only appeared in 1999.
> NeXT was better than OS X Server
How so? UI?
I'm just curious.
> Be was far behind. Looked pretty, but that was about it.
I reviewed it then:
https://archive.org/details/PersonalComputerWorldMagazine/PC...
I think it was more than just a pretty face. It was a demo of how computing should be, and showed that everything else is a bloated mess.
It's much worse now, of course.
> Funny thing is I live in a browser and text editor now, and the Be install I ran in 99 would just about do it!
Which was largely my own thought about Haiku b4. It's very close to doing everything I need.
NeXT was just more mature to me at the time. OSXS was trying to do too much from all directions, while I was of the belief that System 8.6.1 was pretty perfect already.
Agree on Be, and I've tried Hauki off and on. It really seems, as I default to using my phone for nearly everything anyway, that a system like that really is ideal at this point. Like a lot of people my day is spent a jump server away, so I haven't cared what I'm using 99 percent of the time.
OK, ISWYM.
I mean, yes, NeXTstep was much more mature. No argument. But OS X Server was aimed at a different job, and a different job to classic MacOS as well.
People did use NeXTstep as a server -- the first ever web server, notably -- but it wasn't a great one and wasn't its intended role.
Classic MacOS was usable as a server too. It seems to be largely forgotten now but the US Army officially moved all its web servers to classic MacOS at one point, because not having any form of shell at all made it largely unhackable. You can't get shell access when there is no shell to get to, and no clever x86 code exploit will help you when none of the servers have an x86 chip in them.
It wasn't a very good server, though, and quietly I think they changed to something else a bit more conventional not long afterwards.
Compared to Classic, I bet OS X Server was a great server OS. ;-)
I think the real purpose of OSXS was to answer the question:
"OK, we've done a ton of work porting NeXTstep to PowerMac and grafting a kinda sorta MacOS-like desktop on the thing, and it's cost us a sh1t-tonne of money and person-hours of work with not much to show for it so far. Is there any way we can start to make some money off this thing before it's ready to ship as a desktop OS?"
The release of BeOS in 1995 pretty much marked the end of desktop operating systems diversity for nearly two decades. Windows 95 had already been released and quickly cornered the market. Although we perhaps have fewer options today than in the 80s and early 90s things are pretty healthy. No one operating system is dominant in the same way. Alongside Windows, ChromeOS and macOS both have significant shares on desktop and for consumers iOS and Android are just as important.
This repo is more up to date - https://github.com/kanjitalk755/macemu Currently recommended builds for macOS: https://www.emaculation.com/forum/viewtopic.php?t=7360
There's also Mac-on-Linux (MOL), which I've been using from ~1999 to run the original OS under Linux on my Mac hardware until I dropped Mac OS completely a few years later. It does require a PPC host. I remember back then MOL was preferable to SheepShaver, but I don't remember why anymore.
[1] https://www.maconlinux.net/
Fun fact! You can use Mac on Linux to run Mac OS on your Nintendo Wii.
How well does that work? I (not a gamer) only recently learned of the Wii U and thought it would be great for that exact thing.
Doesn't seem perfect but this is also an original Wii and not the Wii U:
https://www.youtube.com/watch?v=qhDQicLvHNQ
I'm not sure if MOL will work on the newer kernel used by Wii U linux though.
> I remember back then MOL was preferable to SheepShaver, but I don't remember why anymore.
Speculation: Running on native hardware probably give you better performance, and probably meant that you didn't have to provide a ROM file because it could just read the chip that's actually installed in the machine.
Sheepshaver runs natively on supported Macs under BeOS and you likewise don’t need a ROM file as it uses the one built in to your Mac. The real reason is probably it doesn’t support MacOS past 8.5 really. Not well.
I had MacOS 9, so that, performance, may have been part of it. Seems MOL emulated an MMU. Although, IIRC I ended up disabling virtual memory / swapping inside MacOS, and just give the emulated Mac more RAM, because doing the swapping purely on the Linux side was more performant. I liked the idea that Linux was simply more powerful, although it might have just been the interaction between the two caches. :)
I remember I was also very pleased to get copy-paste working between the systems, but I may have had to install something for this.
SheepShaver does not run MacOS X, just the classic, while MOL runs both. Been using MOL about at the same time to check out OS X, which led me buying a Pismo, my first and last Apple machine.
It’s probably worth mentioning that you can also use QEMU, which is a bit more involved but can use a bigger disk and also run OS X, and Mac OS 9.2.2 (as it supports an MMU, SheepShaver doesn’t)
Is QEMU the best option to run older OS X (in particular Snow Leopard, and yes, a legal retail purchased physical copy) on a modern machine? I currently keep a rather bulky cheese grater to run a few programs from time to time.
That's the option with the least headache. I believe there are other simulators too but they are much more specialized.
With some hacking you can run all MacOS/MacOS X versions on x86, x64 and PPC from 8.6 to 13 if you have a x64 machine. If you have an Apple Silicon one it gets too slow to simulate around 10.10 (on x64 you can use hvf virtualization). It might be a bit less nice than tailored emulators like SheepShaver but having the same UI for all these OSes (plus other OSes) beats everything in my opinion.
Unfortunately Apple Silicon machines cannot be simulated without some hidden knowledge only some secretive companies have. Hopefully the knowledge will become public some years from now.
QEMU will work the best, but Snow Leopord prohibits being virtualized via its EULA. Snow Leopard Server is licensed to run in virtualization.
My memory is that it only had to be run on Apple hardware, but perhaps I should go back and check. Or just not tell anybody, and hope the Apple Police don't get me — it's not like I'm letting something that old out to play on the internet (though I do sometimes miss iCab).
The only reason I know this is that VMWare desktop products will refuse to run it for this reason. QEMU has no such self-imposed limitations...
Unfortunately qemu isn't as 'integrated' as Sheepshaver -- theres the mouse grab problem, also no disk sharing with the host (the read[/write] fat32 disk don't seem to work, at least under 9.2.2)
I had a a project for #marchintosh of making myself a nice powerPC emulation stack to recompile and open source quite a few free/shareware I did back then, but I more or less gave up as I couldn't find something 'comfortable' to work in. I guess I'll have to find, dust up and try to hook that old G4 Mac Mini that I'm sure is in a drawer, somewhere :-)
You can, but not under BeOS. Under BeOS Mac OS is run almost like Classic was on MacOS X. The OS is more or less virtualised as a user space app, and the CPU is not emulated in any way. That’s the main reason that BeOS only supports PowerPC. Indeed, if you run it on a Mac under BeOS it doesn’t even need a toolbox ROM
Rather a shame that Executor got lost in the shuffle:
https://github.com/ctm/executor
and didn't stick around until Rosetta and so forth went away and it became potentially more useful.
SheepShaver was the successor to the Amiga's ShapeShifter Mac emulator (also written by Christian Bauer). I had fun running Mac OS and games on my Amiga. It also had the interesting property that my Amiga had a 68060 processor, which no Mac had, meaning that it was faster at running 68k apps than any official hardware was.
There is no ppc970 emulator that I know of that can run Mac OS X 10.5 with "Quad Core" PowerPC (64-bit) 16 GB RAM specs. https://everymac.com/systems/apple/powermac_g5/index-powerma...
qemu-system-ppc (2019) can run Mac OS X 10.5 PowerPC https://wiki.qemu.org/Documentation/Platforms/PowerPC
pearpc (2015) can run Mac OS X 10.3 PowerPC with caveats https://pearpc.sourceforge.net/
SheepShaver can run Mac OS 9.0.4 PowerPC
Basilisk II can run Mac OS 8.1 68k
There was the Floodgap guy who got Macos running on the Power9 desktop boards.
I wonder where that went
https://www.talospace.com/2018/08/making-your-talos-ii-into-...
I used Sheepshaver to emulate system non-PowerPC games + System 7.5.3 when making Mac game nostalgia videos on my YouTube many years back. I found it to be far more performant than Basilisk for anything that wanted to run with Quickdraw. It also seemed to have less clipping audio.
I should see how it runs all these years later on a Linux box.
BeOS! Good memories. I tried out BeOS maybe 20 years ago until it was dropped. Is it still maintained?
Haiku is an open source BeOS clone that is still developed, but the original isn’t
Thank you!
You might enjoy this piece I wrote at the start of the year.
https://www.theregister.com/2023/01/11/haiku_beta_4/
You can use haiku os it’s the open source beos and even runs the binaries
But this requires a PowerPC processor under BeOS so you can’t use this with Haiku at the moment. The Linux version looks like it emulates the CPU so it can probably be made to work I guess?
>But this requires a PowerPC processor under BeOS so you can’t use this with Haiku at the moment.
Bold of you to assume a forum full of tech enthusiasts don't have PPC machines sitting around.
No, not really. It isn’t just PowerPC, it is a very specific and picky Mac, has PCI, is old world, has a 603 or 604 processor when shipped originally. If you have one, you then need a copy of BeOS R4.5 or R5, and getting that installed can actually be non trivial (as ActionRetro proved on YouTube - it was my custom pre built image that he finally got working.)
I have 2 BeOS PowerPC boxes, so yep - know all the hassles first hand.
If you do too - awesome, you should come find me on Haiku forums and we can contribute to the BeOS PowerPC preservation effort I started last year.
PowerPC was during my early childhood but our lines never met, so I would be one case where your assumption will unfortunately fall short.
By the time my family got a home computer it was a 120 Mhz Pentium system, but I spent basically all of elementary, middle and high school using m68k and PPC macs at school. During the pandemic i picked up two m68k Macs and 2 PPC Macs to satisfy my nostalgia. I enjoy playing with old tech, but theres nothing wrong if you don't! Different strokes and all that.
Ah, please don't get me wrong: I love playing with old tech too. The first computer I used was some 386 or 486 machine running Windows 3.11, it even had a turbo button and a 5.25" floppy drive. Really wish we still had that thing.
It's just my path never crossed PowerPC's, so I don't have the nostalgia for them that I do for classic x86 hardware. Regardless I'm sure they are an important piece of computing history.
Will BeOS PPC run on a POWER workstation like the Talos Raptor? That would be a hell of a combo...
Near as I can tell the POWER9 chips (at least the IBM ones) only have backwards compatibility through POWER7 from 2010. The last PPC chip Apple used, the G5, implemented the POWER4 ISA, so that's pretty far back.
Nope. It really never even supported G3 officially. It will boot on the BeBox natively an a Mac with a boot loader. Without modifying the kernel, I doubt it will boot on much else.
Haiku should be portable to it, but that's a lot of computer for an OS that is not daily-driver grade.
Haiku will never support legacy BeOS PowerPC apps though, as that makes no sense. PEF is hard to support even today.
It’s a shame, but it’d probably be easier in a VM.
There is also currently no way to even virtualize BeOS for PowerPC. I wish there was. The closest I have seen is a MESS that could boot and show the boot splash screen (this is a 3d Be logo on PowerPC, the one you see on Intel was only ever on Intel.) The dingusppc have BeOS as a possible target for their Mac emulation, but it is not there yet. I'm pretty sure QEMU can't boot BeOS because BeOS requires an old world ROM and it requires a PCI Mac with a 603 or 604 processor - so a 4400, 7300, 8500/8600, 9500/9600 and a few others (including clones) work - but nothing NUBUS and nothing post the first motherboard revision of the 8600/9600. The later x600's have a different motherboard and do not work. Nothing that shipped with a G3 and nothing with new world ROM.
Instead of emulation like this (which is really cool, BTW), I'd love to see something like GNUStep or Etoile growing into a full implementation of all macOS APIs so that I could just build a native Mac application against it and have it run on Linux.
I think that it's often underestimated just how much of what made/makes OS X/macOS good is actually in AppKit and the various other frameworks it's used with. Parts of it are a little aged compared to something that's received more attention in the past decade (like UIKit), but it still does a LOT of heavy lifting and makes developing a polished, capable desktop app much less of a headache.
In writing apps with other desktop UI frameworks, it's become evident just how much AppKit does well that many frameworks barely do or don't offer at all. Sometimes it's even really basic things, like how WinUI somehow has no first-party tableview despite that being one of the most fundamental desktop widgets.
Cross-platform AppKit would be a dream setup.
Indeed. It’d need to be extended to cover the various other subsystems and probably offer a less idiosyncratic look and feel.
Not that I dislike the NeXT visuals - I’d totally have a NeXT as my daily driver (with X, of course), but it became dated the moment Windows 95 copied it.
In any case, anything that can be documented can be implemented in a clean room.
Look like a dead project by now, Mr. Bauer haven't been active on GitHub for years. :(
Maybe we should say it's a mature and finished project. The requirements haven't changed in 15 years.
...except it's not. It crashes regularly and has many compatibility issues.
> for BeOS and Linux
And also for OS X, Windows, and BSD.
Great it's open source software in the form of a git repo now.
But for some reason, using the same repository for 3 independent projects horrifies me.
The commits are unavoidably going to be all over the place, and tracking one project's changes hard.
I don't think they're completely independent.
It's called monorepo for a reason.
TIL.
Still, I see more cons than pros in macemu's case.