Hacker News Reader

Comment by kennend3 | Hacker News Reader

This was my personal favourate statement in the article.

"Deciding whether and when to upgrade macOS is one of the more difficult choices we face. If your Mac isn’t capable of running the current release of macOS, or you’re dependent on key hardware or software which is incompatible, then the decision is made for you."

If you head over to the hackintosh side of things you will notice that very rarely does a hackintosh not upgrade to the next version.

I don't want to get "conspiracy theory" here but I do wonder how often the next release is blocked by Apple, not by hardware?

There are hacks to allow the new OS to run on real Apple hardware despite being blocked and people who have done this seem to be running the new OS just fine?

Releasing an OS with serious bugs, then limiting someone from upgrading to a newer one which fixes those issues is just bad.

View on HN

I have a Macbook Pro 2015 and the latest OS it supports officially is Monterey. The new Ventura (2022) is not supported anymore. I consider my device to have had software support for about at least seven years so far and until when they stop releasing security fixes for Monterey. I believe this to be sufficient for my needs and the frequency of which I upgrade the laptop.

P.S. I don’t think that it’s ok for Apple to stop supporting older devices. I am just pointing out my personal view concerning my needs.

My 2015 MBP runs Mojave (~2020) and will until it dies. Catalina was just too much for me.

Besides, OS X looks much better than the newer iOS-based design, imho.

I wish apple would do 'best effort' support on older macs. Meaning, it'll let you install after you click thru two pages of warnings, and might work, or might not.

This is a much more sensible approach vs the "installer lock" apple took.

But it isn’t the way Apple does things. They obviously don’t want Apple computers in the wild running a version of an Apple OS that might or might not work.

Yes.. artificially limiting them from upgrading by checking the SMBIOS in the installer and preventing an upgrade is MUCH BETTER for the users.

And as posted here several times, what about those who use opencore legacy patcher and upgrade despite the block. What "doesn't work" for them?

SO to your point, apple prefers machines in the wild running an old and out of date OS because???

Apple ships all drivers with OS. So, if Apple drops hardware support, it removes drivers and other support for not supported hardware. For example for Ventura they dropped support for Nvidia GPU-s, Metal is built now to require AVX2 and much more. "Best effort" support is not trivial in such cases. Yes, I know about OCLP and use it myself, but it's a hack.

I tell people that I haven't looked forward to new features in the 'California' era like I did in the 'Cat' era. Now, I worry about what Apple will break or take away.

If Microsoft wasn't so intent on spying and advertising to me, I might consider them. I get the feeling I will switch to a BSD eventually.

I’ve been leaning the same way. I’m curious why you also lean towards BSD instead of Linux?

It’s a gut feeling for me after watching the Linux ecosystem churn a lot and seeing BSDs with well documented core systems and the “core/ports” distinction make clear what is and isn’t really supported

I'm much more familiar with OpenBSD and FreeBSD on the server-side, and I don't feel real comfortable with Linux. The two biggest vendors, Red Hat and Ubuntu, don't really give the warm fuzzies. Maybe something like System 76 backs.

What about plain Debian :) if you’re comfortable with BSD then Debian should be a cakewalk for you.

The funny thing is I tried to setup and Arch Linux box. I can setup OpenBSD with thinking with all the partitions and such, but that Arch Linux install was just a bad time. I'll try Debian when I get back to work. We are replacing our Samba server and the BSDs aren't real current given some changes the Samba team made.

I have tried all three and much comfortable with NetBSD.

Well, don't skip the DragonFly folks. I've had the four main one up and running at different times.

There is always OpenDarwin/PureDarwin?

https://en.wikipedia.org/wiki/Darwin_(operating_system)

I prefer BSD to Linux for a few reasons. Firstly, BSD is not encumbered by GPL. Also, any BSD is a complete operating system while Linux requires distros subject to the whims of their creators, which seem to focus on form over function. Any long term stability is almost accidental, and there is no standard edition of Linux. Lastly and I think most importantly, the centralized BSD ports system is more trustworthy than an untracked number of repositories and sources. Actually, these complaints about Linux aren't about Linux, they're mostly about GNU. Linux is just a kernel, and it's fine, really. GNU is the problem.

Agreed. Also all BSDs have different characteristics.

Linux it self can't use GPLv3 shows that it is not innovation supporting License

> Firstly, BSD is not encumbered by GPL. Also, any BSD is a complete operating system while Linux requires distros subject to the whims of their creators, which seem to focus on form over function

This "BSD is a complete operating system" that you speak of... Do you mean FreeBSD, OpenBSD, NetBSD, DragnonFly BSD, Darwin?

Thankfully BSD is not subject to teh whims of any distro?

Each of {Free,Open,Net,DragonFly}BSD is its own operating system, complete with tree, system call interface and ports tree. This is not the same as the distribution model Linux has - the BSDs have diverged substantially from one another.

Has anything workable been made from Darwin in the past decade or has Apple mostly left that community high and dry?

Coming from macOS, there’s a certain cohesiveness that one comes to expect that is often absent with Linux, with conventions being mere suggestions and things generally being all over the place. The BSDs aren’t perfect but certainly much better in this regard.

It’s unrealistic but a Linux distro that forked as many packages as possible to make them follow conventions and be proper desktop citizens would be interesting.

Check out the KDE edition of Fedora. Fedora is a rock solid OS and KDE is a rock solid desktop environment. Reminds me of what I always wanted Windows to be.

You can go with their default edition (GNOME) if you want something that looks like a MacOS clone right out of the box, but the actual user experience is so different (and IMO so much worse) that I can't recommend it for anyone interested in switching.

Funnily enough I've been considering going the opposite way. I'm on Windows, Android and iPadOS right now, but this diversity of operating systems means nothing integrates well, so it's not a great experience.

Buying into the Google or Microsoft ecosystems would improve things on this front, but this seems not great w.r.t. privacy, which is why I'm thinking of going all in on Apple.

There is no excuse for the most profitable company in the world to drop support for working hardware. It's entirely within their capability to support even PowerPC macs. And that's besides fact that Apple has an incredibly small portfolio of hardware to support; so much so that they could trivially exhaustively test every update on every possible piece of hardware.

I was just rereading the ppc history for Apple as they're doing it again for ARM and it seems quite reasonable to be honest. They pulled something off I think would be difficult for many others and have shown they're not married to a single architecture.

Still RIP my favorite machine the last ppc Mac tower

> They pulled something off I think would be difficult for many others and have shown they're not married to a single architecture.

You mean like how MS Windows ran on the DEC Alpha as well as X86? Or how Linux runs on X86 and ARM CPU's?

Running an OS on multiple CPU types goes back a very long time.

> Running an OS on multiple CPU types goes back a very long time.

Yet moving the entire world onto this new architecture and getting all developers on board to start creating software for it is nearly unheard of outside of Apple.

That’s the strategy that Microsoft can’t/won’t pursue: be willing to abandon legacy vendors/customers by forcing a migration.

I’m fan of how Apple does things for their ecosystem. I think only this strategy could have let them make the transition to arm as smooth as it’s been.

That said, I also have deep respect for Microsoft for being so deeply committed to backwards compatibility. You can find developer stories about how they’ve gone out of their way to hard code support for legacy apps (even space pinball) in major Windows updates. That sounds really gross from a code perspective, but from a business/product perspective, it buys trust/loyalty from enterprise customers. We see the opposite issue with Google, who has established a reputation for unreliable long term support. Despite their reputation for strong engineering, their new product lines suffer from distrust. Potential customers ask themselves why pay the onboarding cost for a Google product when it’s unclear if it’ll be around in two to three years.

"Planned obsolescence" has its advantages?

This is the third time apple has done this so one would hope they have learned form their mistakes and make the transition seemless?

Windows machines on X86 are just fine.. Amazing when you dont have a monopoly on both the OS and Hardware. MS runs on Intel, AMD, ARM...

MacOS runs on whatever hardware Apple sells.

you can see how an attempt by Microsoft to move off Intel would fail, right?

You mean like how MS Windows ran on the DEC Alpha as well as X86? Or how Linux runs on X86 and ARM CPU's?

It’s not the same thing.

A macOS hard drive could boot either an Intel or PPC Mac and all of the apps ran on both types of machines.

Apps that were “fat binaries” ran natively on both processor architectures.

Same thing today with the apps and the operating system running natively on Intel and ARM machines.

Nothing you wrote is accurate.

The apps ran on both types of machines thanks to rosetta and the performance hit this takes: https://en.wikipedia.org/wiki/Rosetta_(software)

It is doubtful that a PPC system can boot an intel MacOS HD. One is RISC the other CISC.

The reason you can "natively" run intel on ARM machines is Rosetta2

"Until recently, Mac developers only had to worry about creating Intel-based apps. However, that started to change when the first Apple silicon Macs arrived. For Intel-based apps to run on these computers, Apple introduced Rosetta 2."

Please read up before posting.

lol maybe you should do some reading before posting. Mach-O supports multi-architecture binaries. Rosetta is solely for things that haven’t been made “Universal”

https://en.m.wikipedia.org/wiki/Universal_binary

Great.. now explain why Rosetta exist if it is not needed because "mach-o supports.."

> It is doubtful that a PPC system can boot an intel MacOS HD. One is RISC the other CISC.

That's the point of fat binaries: they contain .text sections for several architectures. Even the installation DVD was a single one, for both platforms.

Some more modern example of the same concept:

    Terminal: Mach-O universal binary with 2 architectures: [x86_64:Mach-O 64-bit executable x86_64] [arm64e]
    Terminal (for architecture x86_64): Mach-O 64-bit executable x86_64
    Terminal (for architecture arm64e): Mach-O 64-bit executable arm64e

Nothing you wrote is accurate.

You must be new here. ;-)

The apps ran on both types of machines thanks to rosetta and the performance hit this takes: https://en.wikipedia.org/wiki/Rosetta_(software) It is doubtful that a PPC system can boot an intel MacOS HD. One is RISC the other CISC.

I know about this because I used to do it all the time when I worked in tech support. Regardless of which Mac a client might show up with, I could boot it with the same external hard drive back in the mid-90s.

Rosetta allowed unmodified PPC apps to run on Intel Macs by translating them on the fly.

However, developers had the option of making a few changes and recompiling their apps to produce a version that would include native code for PPC and Intel machines; macOS would handle running the correct binary. These were known as "fat binaries".

Fast forward to 2020 and Apple is applying the same strategy. Rosetta2 translates Intel apps on the fly to run on ARM. And developers can recompile their apps so that they run natively on Intel and ARM machines.

Even though I'm typing this on an Intel Mac, I could boot an ARM-based Mac with its hard drive if necessary.

Here's the output from the file command on the ls command:

    /bin/ls: Mach-O universal binary with 2 architectures: [x86_64:Mach-O 64-bit executable x86_64] [arm64e:Mach-O 64-bit executable arm64e]

    /bin/ls (for architecture x86_64):      Mach-O 64-bit executable x86_64
    /bin/ls (for architecture arm64e):      Mach-O 64-bit executable arm64e

Yes, even ls is a fat binary that runs natively on Intel and ARM-based Macs.

Please read up before posting.

Perhaps you should take your own advice?

Well, you can't run 2002 OS X software on a modern Mac anymore, but you can still run 2002 Windows software on a modern PC (and probably 2002 Linux software too).

Actually, you can't even run 2010 iPhone OS software on a modern iPhone anymore (but you can run 2010 Android software on a modern Android phone).

That's not true in any practical sense.

Sure you can run the xp calculator app on windows 11, but so what? Most of the programs that people want backwards compatibility for, enterprise apps, won't run correctly. Yes that's mostly because of how they're coded, but even if the source code still exists, I doubt anyone is putting effort into updating them at this point, and is just running xp or 2000 in a vm.

This is sort of my line of thinking. when you effectively have a very small product line and are incredibly profitable it does seem a bit greedy to artificially limit updates.

It would be different if there was no way to bypass the upgrade lock, but given opencore legacy works.. Clearly the hardware is able to run the new OS.

Besides the impracticality of running new OS’s on a 2005 era Mac, what software engineer at Apple would be willing to take on that role? Also, even if the hardware were supported, what third party software company would still support PPC Macs?

> what software engineer at Apple would be willing to take on that role?

One being paid to do so? People still work with COBOL. If you have trouble finding engineers to do it, offer a higher salary.

> Also, even if the hardware were supported, what third party software company would still support PPC Macs?

Some software is still supported for Windows XP, and that's considerably older than the last PPC Macs. It would certainly be a whole lot easier to support if Apple still did the same.

In any case my point is that they absolutely have the resources to do this. I don't see how that's at all controversial given how much money they have.

There's still Linux distros supporting 32bit PPC Macs.

And how many people are actually getting paid to do so and how many companies are selling software for them?

> One being paid to do so? People still work with COBOL. If you have trouble finding engineers to do it, offer a higher salary.

Higher salary than who? Do you really think old COBOL programmers (FWIW I’m 48) are making more than the average new college grad working at any major tech company?

And besides if you know how raises and promotions work at tech companies, you would know they are based on “impact” and “scope”. You don’t demonstrate either by saying “I spent all last year supporting MacOS on a PowerPC 6100 from 1994 doesn’t speak well to either.

Unless you work at Google where it’s based on creating new messaging app that will be killed in a year.

Is the average COBOL programmer even making close to the same $170K+ that I’ve seen returning interns make at major tech companies (including the one I mentored)

> Some software is still supported for Windows XP, and that's considerably older than the last PPC Macs.

How much mainstream software is supported by XP? Microsoft Office? Adobe software? Modern browsers? Modern games?

> In any case my point is that they absolutely have the resources to do this. I don't see how that's at all controversial given how much money they have

Yes and they also had the money to buy Twitter. But what moron would do something that stupid?

We can indeed run newer macOS on older macs with the Opencore Legacy Patcher - https://dortania.github.io/OpenCore-Legacy-Patcher/ . But the bugs aren't the only concern. There's also the revised terms of services and privacy policies that we have to consider with each new macOS, as Apple plans to become a service company at our expense. It is also clear that Apple has been taking away developer options, and rearchitecting macOS to be more and more like ios / iPadOS to have greater control of over what developers can do on the system. (There's now even talks of Apple working on a hybrid macOS and iPadOS merge so that they have one common OS for both platform).

Reply on news.ycombinator.com