Apple Announces Marzipan for 2019
Mark Gurman (tweet):
During its annual developer conference, the Cupertino, California-based company said the two app platforms will merge next year, allowing apps built for its iOS mobile operating system to also run on Mac computers. Bloomberg reported the planned change in December, but Apple hasn’t talked about it publicly until Monday, when it gave a preview.
The point of this is not to create a single unified OS, Federighi said. But the fact that Apple spoke openly about an initiative that could arrive as late as a year from now is a clear nod to how the tech giant perceives the future of apps.
At WWDC on Monday, Apple said some of its homegrown iOS apps, including Home, Stocks, News, and Voice Memos, would be available later this year on macOS Mojave, the next version of macOS. On the surface, it seems obvious that Apple might make some of its own apps available across different platforms. But behind the scenes, Apple has been building tools third-party developers can eventually use to port their own apps from iOS to MacOS with what Federighi insists will be minimal effort. Apple began this initiative around two years ago, and its own internal software engineers have been beta testing the tools. These four iOS apps for Mac are the products of that testing.
Still very light on technical details. I get why they announced this a year early — because they wanted to start shipping their own apps built on this — but it’s so unusual for Apple.
Craig says the goal is not to merge iOS and macOS, but that is inevitably where this will end up.
“We’re not merging macOS and iOS!” …but btw almost all the new Mac apps we showed today came from iOS and use UIKit
The ARM transition seems so obvious at this point, it’s not even news. But UIKit on macOS was definitely not how I expected them to do it
Marzipan is wild. It’s running the UIKit apps with XPC & remote rendering. Not quite Mac OS Classic.app, but it totally feels like an alien OS hiding in the background
Oh man
So, Mac developers, how are you feeling about UIKit on macOS now you’ve had time to digest the news?
Right now, it seems all-or-nothing; I would have expected apps with both AppKit & UIKit code, interleaved views, etc — but no way if they have separate window servers like now. That may change over timeSteve Troughton-Smith added,
This is a fun UI on macOS
The current approach isn’t really evolvable into hybrid AppKit/UIKit apps at all so I’m not sure where this goes. The way the stack is oriented it’s not even remotely positioned to replace AppKit in the stack. Good for us AppKit lovers I suppose but generally confusing.
Ok one more Marzipan hot take.
It bothers me that if Marzipan is the “so you can’t afford to make a real Mac app” stopgap that Apple would ship key built in apps with it.
I know it’s dogfooding but it’s also Apple putting out second class Mac apps by their own definition.
From how I interpreted @stroughtonsmith’s spelunking so far, it is somewhere between a VM and NSOpenPanel. Out-of-process but transparent to the client app.
I think they’re taking the opportunity to move more of their code into a micro-kernel-like silo.
UIKit is not young. It’s about as old as Toolbox was when discussions started about it’s future, and it also is on a platform that’s outgrowing it.
Sticking it in a virtual environment is the perfect way to not commit to it while deciding your next move.
Previously:
- Apple Rumored to Combine iPhone, iPad, and Mac Apps to Create One User Experience
- Tim Cook Says Users Don’t Want iOS to Merge With macOS
- Scuttlebutt Regarding Apple’s Cross-Platform UI Project
Update (2018-06-06): Jason Snell:
It feels like Federighi’s cutting it awfully close on the “unique ergonomics” front, though. Apple itself sells a keyboard for the iPad Pro, and even allows users to move their fingers like they’re using a trackpad when editing text. The iPhone and iMac Pro are about as far as two devices can get from one another, but the MacBook and the iPad Pro are not.
In his interview with Wired, Federighi also pooh-poohed the idea that Apple might make a MacBook with a touchscreen. And yet one of Apple’s greatest arguments against touch on macOS—that Mac software was designed with keyboard and pointing device in mind—is going to very rapidly become obsolete as iOS-sourced apps appear on the Mac in 2019 and 2020.
I think it’s time for a new word…
Marzipandemic: The spread of single-window odd-UI originally-on-iOS apps to the Mac.
Update (2018-06-07): Colin Cornaby:
I still really wish they’d actually made a UIFoundation that was shared between Mac and iOS and had common elements. It would solve all this...
But Apple, even publicly, is pitching this as a repackaging process of existing iOS apps, not developing Mac first apps with UIKit.
Although I want to believe Marzipan will kill Electron, I doubt it will. Electron isn’t used by iOS developers who don’t know how to make a Mac app. It’s used by web developers who don’t know how to make a Mac app.
Update (2018-06-08): Steve Troughton-Smith:
Developer/user question: what features or UI elements does a Mac app need to have to be considered a good platform citizen? i.e. in your opinion what specific things does UIKit/Marzipan need to grow to actually be viable for Mac apps?
Update (2018-06-11): Jean-Louis Gassée:
In the end, Apple’s UIKit port to macOS raises questions we can’t answer. iOS apps such as News, Stocks, and Home ported to macOS are nice, but, again, I don’t see a stampede of iOS apps crossing the bridge to macOS, not enough to move the Mac volume needle. This leaves us with two possibilities. Either the UIKit move is a titillating but unimportant sideshow, or it’s part of a larger hardware plan for the Mac.
See also: Pierre Lebeaupin.
Update (2018-06-11): Mike Rundle:
Apple getting rid of subpixel antialiasing in Mojave while also applying a non-integer scale to Marzipan UI elements is some kind of sick joke.
Update (2018-06-13): Lauren Goode:
In other words, Frankensoftware might seem like the wretched experiment of a bunch of FOMO-driven executives when you’re struggling to swipe, tap, or shout your way through an interaction with a new product. But from now on, most every connected thing you buy is going to have a little bit of something else in it. And once the companies making those things figure out a way to make these interactions effortless, it won’t seem like such a bad thing.
Update (2018-06-14): Guilherme Rambo:
Any questions about iOS apps on the Mac? Ask me.
Update (2018-06-18): Chris Adamson:
That’s something that makes me skeptical about the Marzipan dream. Developers who give a crap are already able to make this kind of thing work, and developers who don’t give a crap aren’t going to do it no matter how easy you make it.
Marzipan (candy frosting) is a legacy porting technology: Existing iOS apps can cost more to port to AppKit than they’re worth, but may be worth something as a cheap Marzipan port. Nobody ports their iOS apps to tvOS or watchOS because it’s not profitable, and everyone (in the first world with money) has an iPhone already.
Update (2018-07-05): See also: Rob Fahrni.
Update (2018-07-11): See also: Accidental Tech Podcast.
Update (2018-09-03): Melbourne CocoaHeads:
New video from July 2018 - @argentumko presenting Mac OS Mojave Marzipan’s Internals.
I know it’s not done, that it’s a really hard problem, and that it’s “handy” to have Home etc in Mojave, but it’s all just so wrong.
Remember how we all laughed and laughed at Microsoft because Windows 8 had “tablet” apps & “native”/old apps?
How is this better/different?
These apps are not good. It’s that simple.
Update (2018-09-04): Andy Ihnatko:
It kind of pisses me off. Because these system apps are also supposed to show developers and users how good these kinds of apps can be. If Apple’s own engineers can’t hit this pitch, it’s a missed opportunity at best and a bad sign at worst.
…but they are ‘good enough’, and better than the Mac missing out completely on a viable app-led future. Such is the story of Marzipan
They’re good enough to launch with, but they’re clearly not the final destination. Initially they’re going to be like every other platform Apple has done. The “defaults” will get developers 90% of the way, and the last 10% is what will make certain apps shine.
“Insanely good enough!”
I agree with the first half. I want UIKit to take over as the UI framework on Mac.
However, I don’t want iOS UI elements on the Mac!
I SERIOUSLY hope the current UIKit on Mac does not become the ‘new normal’. That’d be the start of the death of the Mac.
The ‘start of the death of the Mac’ was ten years ago. This is the start of a new chapter of life for the Mac
My position here is that Apple just needs to do an outstanding job. Devs are gonna do what devs are gonna do, but if Apple makes it easy to make crap, we’ll get lots of crap.
Apple, and the community, needs a definitive takedown on how we waited a decade for Mac-like apps on our iPads, and ended up with ugly iPad apps on undependable, neglected Macs.
The current Mojave apps feel like I’m using the iOS Simulator. If all Mac apps were like this today I’d be a Windows person 100%.
The other thing that always seems to get missed is that the apps that Marzipan can bring to the Mac and the apps that are missing from the Mac are two different sets of apps.
Will Marzipan bring Windows only software to the Mac? No? Then what apps are we talking about?
“But Colin you’ll be able to run Yelp”
Yeah seems to be running just fine already.
Update (2018-09-06): Peter Steinberger:
Here’s the slides for my Hacking Marzipan talk + lots of links to learn more[…]
4 Comments RSS · Twitter
[…] Previously: Intel and the Danger of Integration, Apple Announces Marzipan for 2019. […]
[…] Apple Announces Marzipan for 2019, macOS Mojave: Back to the Mac, Tim Cook Says Users Don’t Want iOS to Merge With […]