Slashdot Mirror
Apple To Phase Out 32-Bit Mac Apps Starting In January 2018 (macrumors.com)
Apple will be phasing out 32-bit apps with iOS 11, and soon the company will make the same changes on its macOS operating system. During its Platform State of the Union keynote at the Worldwide Developers Conference, Apple told developers that macOS High Sierra will be the "last macOS release to support 32-bit apps without compromises." MacRumors reports: Starting in January of 2018, all new apps submitted to the Mac App Store must be 64-bit, and all apps and app updates submitted must be 64-bit by June 2018. With the next version of macOS after High Sierra, Apple will begin "aggressively" warning users about 32-bit apps before eventually phasing them out all together. In iOS 11, 32-bit apps cannot be installed or launched. Attempting to open a non-supported 32-bit app gives a message notifying users that the app needs to be updated before it can run on iOS 11. Prior to phasing out 32-bit apps on iOS 11, Apple gave both end users and developers several warnings, and the company says it will follow the same path for the macOS operating system.
Also it means I have to throw away my Core Duo based macs. This is a bit why, after using macs and OSX for 15 years that I have walked away from them. I can't keep up with the pace that Apple has set for consumers.
“Common sense is not so common.” — Voltaire
How horrible Apple is for offering all those free updates (including the OS) as long as the hardware supports them. Rebooting after the update is so much trouble. Remind me which company you work for please, so I don't ever get a job there having to support that 10-20 year old hardware.
And why is that crazy? If the software runs on a standalone computer, not connected to the internet, why does it matter? As long as it runs and does what it has to do, what's the problem?
Well in my opinion, the AC you're responding to made a bad comparison. What he's complaining about is not really like a business building a physical plant or factory and expecting it to run for 50 years. It's more like a business building a refinery, and expecting it to run for 50 years without any maintenance, upgrades, or cleaning, meanwhile firing everyone who understands how the refinery works or knows why it was built the way it was.
To answer your question more directly, there are a few different problems with expecting to run a legacy piece of software on a standalone computer, not connected to the Internet, for decades on end. First, if the system needs to interact with other systems, then you can't really isolate it from the Internet. It might not be directly on the Internet, but it could still be compromised from another computer that is connected to the Internet. Second, what if the hardware the software runs on breaks? Are you going to be able to get replacement parts in 20 years?
There's also another problem that's a little bit more nebulous and therefore harder to argue for, but: You don't know what you're going to need in 10 years. I've had to deal with companies that have some old unmaintained business-critical application running on a server somewhere. Since they began using this application, they've upgraded all their computers 3 times, and started using a couple other applications. They'd like to have their various applications talk to each other instead of doing manual entry to keep the apps in sync, but they can't, because the old app wasn't designed to do that.
They want to be able to access the old database from an iPad, but they can't, since it requires a Windows client-side app. It might be that they've been continuing to buy Windows 7 machines for the past several years because the client-side app doesn't run on newer versions of Windows, and they don't have a way to update it. Once Microsoft started trying to force everyone to use Windows 10, this has become a problem. Maybe they'll have to run a Windows 7 VM if they ever want to update their desktops again.
Then they go to update their networking equipment, and they want to change their IP scheme and move their servers to a new VLAN. They can't do that. Whoever wrote the application seems to have hard-coded an IP address somewhere, and changing the IP address of the server breaks everything.
The server they're running the application on developed a problem where it crashes occasionally. The IT team realizes that the hardware might be failing, so they want to spin up a VM on a new server and install the app from scratch. They can't. Nobody at the company knows how to install the application anymore. They instead P2V the existing server, but the problem still occurs. Oh well, I guess they'll have to live with occasional crashes.
I'm kind of cobbling together different examples of problems that I've run into over the years, but the point is, this crap always seems to turn into a mess. If your business relies on a piece of software, you should have some support contract with a developer capable of supporting and (if needed) updating the application, installing it from scratch on a new computer with a brand new OS.
32-bit ios actually stands to be the first "lost platform" in computing history. EVER!
Hyperbole? Hear me out:
Unlike almost every other platform, there's no reliable and good way to run ios software (or ios itself) outside of the hardware. The only things that look like emulators are open source, and you can't even choose to install older versions of the OS on hardware past a cut-off date. Apple has fully DRMed their platform, fully closed it off. But up until now, if you have played by their rules, there's always been a way to run any given application: the expectation that your app can't be emulated well on Linux or whatever isn't something universal, so the computing consumer world has been pretty accepting of this.
This fully closed and cryptographically sealed system is something reasonably new in computing, and Apple's smashing success with it has encouraged others to duplicate it to some lesser degree- Windows 10S tries to take their model and offer a greater degree of freedom with an opt-out for cash (instead of no opt-out), Android has taken parts of their system, etc. But so far, everything has eventually (once it is no longer a primary economic driver) been emulated, been archived, been available for the future. Perhaps the loss of 32-bit ARM code compiled for ios is no great eternal loss to the world, but the precedent is now set for the expiry of code in a way that has never been done before.
AMD64 was introduced back in 2000.
You know, long before Moore's law became bow-legged from the heavy burden of asterisks. (Yes, like always before, we do indeed have more transistors, but just try to use them all at the same time and see what happens ...)
So that's seventeen years ago. Subtract another seventeen years, and we're back to 1983.
Back in 2000, your karmic twin would have been moaning about the loss of 8-bit software compatibility.
Subtract another seventeen years, and we're back to 1965.
Back in 1983, your karmic triplet would have been moaning about the loss of slide rules.
Lament for the Slide Rule — August 1985
Unfortunately, that's paywalled, so we're stuck with this belated cuckoo:
When Slide Rules Ruled — Cliff Stoll (2006)
Check out this giant pull-quote:
Nice. That saves me from craning my neck to look through my window for plummeting petunias. You just never know anything with absolute certainty.