I like the concepts presented here but I cannot shake off the feeling that OS X is pretty much never going to get the love it deserves from Apple.
Apple will just keep improving iOS to be more powerful, more flexible and so on. The iPad Pro with iOS 9 is just a rough draft, it will be iOS 10 or 11 that'll show the hardware off, at least I hope. iOS 9 was probably the biggest jump in productivity for the iPads and hopefully, Apple keeps it going from now on, not neglecting it like it did prior to iOS 9.
Well the Mac has had a lot of love poured into it over all these years to get it to where it is today. I enjoy using the Mac in its present incarnation immensely.
El Capitan was that for me.... Mavericks and Yosemite were all kinds of buggy, to the point where I was rebooting once a week due to hard-to-pin-down slowdowns, but at the last time I rebooted El Capitan due to an OS update, I had 40 days of uptime.
I agree with you that Apple probably sees iOS as their future. I am not sure that I care too much. I do have and old Macbook Air and an even older Macbook Pro that I get a lot of service out of, but I am also good with using Linux on two other laptops.
For developers I think the future is cloud based web dev setups anyway. I really like kicking off a large build or run on a remote server and not have it heat up my laptop. For different purposes, having a few beefed-up VPSs available and services like nitrous is more convenient. That said, developing in Emacs or vi in SSH term windows lacks some convenience, so it having deep integration of IDEs like IntelliJ with your own server build processes would be interesting (and even more interesting to also have great iOS and Android support).
EDIT: more on the topic of OSX: I much prefer the small continuous update process that OSX and Windows 10 seem to be taking. I use Ubunutu and I often think that I would prefer small continuous updates than 6 month drops - at least for laptop use.
Tablets and smartphones have a touch screen, computers most often don't. Software that doesn't take this into account will be a second-class citizen on the respective platform.
And while a flat filesystem with a powerful search function certainly has its uses, I wouldn't want to use it for everything. I'm far too anal when it comes to organizing my source code, for example.
> "And while a flat filesystem with a powerful search function certainly has its uses, I wouldn't want to use it for everything. I'm far too anal when it comes to organizing my source code, for example."
The article touches on a concept called 'loops' for cases such as this.
I think there's room for innovation with filesystems. Personally I'm a fan of the 'database filesystem' ideas from BeOS/Haiku.
> Personally I'm a fan of the 'database filesystem' ideas from BeOS/Haiku
How do you see that working well?
Microsoft tried to incorporate all the stuff like MP3 metadata into explorer way back when (XP? Vista? Can't recall)
OS X has supported "saved searches" which act like a live database-backed file manager ever since Spotlight was added. More recently they added quick tagging of files.
What are you imagining that would be more usable than these?
I have my music, my music app queries the database, presents my with my music, tags them into playlists etc, great.
I have my photos. OK, so I want to keep my screenshots and the memes I downloaded from imgur separate, but those are fairly easy to special-case.
Now I'm writing a novel in my spare time. I got my text document, I got whatever Evernote/OneNote-style system for notes. I found this cool MMORPG map editor I'm using to keep track of where my characters live and go to work, does that count as a picture. And I recorded when I hummed a few bars because I wanted to add a song my character wrote when he was bored, and now it keeps shuffling that in my music. Hmm these things are getting difficult to associate now...
I'm not familiar with those other approaches, but customisable metadata is only half what makes the database filesystem beneficial. Can you specify folders using a query language (via a GUI or otherwise)?
... Except that BeOS didn't really put anything that isn't also present in Windows or (much later) in OSX, as well as at least some implementations in Linux.
And I might be a bit pedantic, but filesystems are databases. Indexing and navigation methods sometimes differ, but a lot of people use advanced indexing in windows that not only checks extended attributes but also contents. etc.
These concepts look nice, but never take into account the difficulty of working on complex projects with multiple people that hierarchical filesystem actually work well on.
There's absolutely no way you could build an app in a big "bucket 'o' files". Sure, you can put together a flyer for a softball game or (maybe) type up a simplistic report for school, but eventually these "loops" are going to become big, unusable buckets of content that isn't easily searchable (do we have good search on music contents yet? better hope the filename is good before it goes into the bucket!).
Perhaps the answer here is a "subloop". That's starting to look like we've come around to nested document folders again, isn't it?
Somewhat amusing: the screenshot of Sublime Text in the mockup is using folders.
What I find most funny about it is that HFS actually provides the closest model framework for a heirarchial distributed FS, where pretty much no other does (btrfs does have some benefits as well.)
Like the author, I worked on a distributed cloud FS, but hierarchy was a core goal. The abstracted reference system of HFS that split meta-data from the data resources is a huge boon to handling various common issues, from rights management to doing "fast" full-tree cloning over a distributed store and several other neat tricks like persistence. The biggest issue with distributed heirarchial file-systems is probably how to handle the full-tree cloning, in a way that isn't crippling, slow, or causes an enforced depth and file limits.
I would rather have a modern file system optimized for SSDs instead of one with 16bit alignments for the Motorola 68k. It seems like OSX versions are more about tabs in Finder nowadays while Linux (and Windows) actually innovate.
I don't see anything new here. Some apps were redesigned to behave a bit like Photos.app, but pretending that adding a menu bar and multiple window support will be an easy port shows an amazing lack of experience. edit: just realized that there are no multiple window apps. So, the distinction of a computer vs a mobile device is gone right there
Organizing files into projects is waved away by "just search through an amorphous blob of data". "Kind: PDF". Who finds files like that?
The user activity thread is nice, but it's clear Apple was already working in this direction with the links timeline in Safari, but all the social service TOS's forbid this kind of presentation and they're super-protective of their presentation to the point even Apple can't negotiate past it at this point.
I find these sorts of concept look great, and look they would work well, when you have 10 emails in your mailbox, or 3 tabs open in your browser, or 6 files in Sublime Text, etc etc.
I get hundreds of emails a day, work in an editor that is indexing ~11000 files, regularly have 50 tabs open in Chrome, etc. I find many apps that follow the sorts of aesthetics and principles shown in these mockups just do not scale to what I would consider normal professional use.
I think your impression of normal professional use is bonkers. For example, we use Slack and I auto-filter auto-emails so I only get about 30/day in my inbox. While you probably don't need 50 tabs open, I'm not sure this direction would preclude handling that in a decent manner. The editing would work the same whether or not Finder presented the files hierarchically or not...no change there.
The described behaviour isn't that unusual. Several of the people I've worked with have said they do things in almost exactly the same way. (and they're not idiots ;>)
It's not that these things can't be done, but I find time and time again that when you simplify software and focus it on typical consumer usage patterns, it fails to work for professional use.
- A "people focused" email experience doesn't work very well with monitoring emails. I've had to turn off Gmail's smart inbox because me being CC'd into a person to person thread that I don't care much about is "high priority", but that Monit email saying CPU is at 99% on one of the webservers looks like every other Monit alert, and is a low priority update.
- I use Slack extensively at work, but email is still incredibly important.
- I may not need 50 tabs open, but after my 5 pinned tabs, a few reference/documentation tabs, the prod version of what I'm working on, the local version of what I'm working on, the PR page, and a few tabs of personal stuff, I'm easily reaching 20+, all it takes is one context switch to a different project and I've opened another 10. 50 isn't uncommon.
- A hierarchical file system is pretty important when working with Git repos. And I'm not a contractor, but if I were, separating client data is of paramount importance.
I don't get why people want osx to die and ios to run on mac. They are already pretty much the same OS. Open up the OSX UIKit port should be enough to solve that "problem".
Also, removing the file system tree would be a huge step backwards. Organizing hundreds or thousands of files between dozens to hundreds of projects is just not suited for a "everything in one searchable bag" for a few reasons: To find something with search, you have to already know what you're looking for - something that browsing a folder solves instantly (how can you know you see all relevant files with a search? Maybe there's an important file that didn't match your query?!) - and the common "solution" with tagging/labels are just single-level folders in disguise....
I'm glad Apple decided to not port OSX to the iPad, or replace OSX with an iPad-esque iOS. Most people I've talked to that experienced windows 8 would agree that when Microsoft did just that, it was a misstep.
I'm a developer so I live and breath the file system tree. On the other hand, don't forget that the original Mac did not have a file system tree. Rather, it had a bunch of linked objects through which you could navigate. (It also had flags for file revision numbers which was never implemented.)
An article I read way back in the early days of the internet said people were in one of two categories: Searchers and Navigators -- that is, people who were more comfortable getting what they want by searching, and people who were more comfortable getting what they want by navigating a hierarchy.
That really struck me, because I'm of the latter type, and it makes me somewhat uncomfortable to have a search paradigm thrust upon me.
I don't mind searching to find what I want, but I find myself uncomfortable if I can't then see where it "really" is. In a universe with just tags/labels, it doesn't feel like it's anywhere.
In case you hadn't considered the alternative position, I get nervous deciding where something should "Really" be when that's based on a number of things, e.g. there's a multitude of hierarchies I could pick when organizing a music collection, and none is inherently correct over the others. Some things defy a hierarchical structure entirely. (I hate having a "Various Artists" folder, and those albums not being visible under each artist. I have considered symlinks, but too often software treats these as multiple copies.)
I can avoid that by having my files in a content-based-addressing pool with metainformation.
Trees and hashes both scale to any size, it's about their performance based on how you access and modify them that matters.
Some collections of my files are accessed entirely hierarchically, others are not, I don't think I should have to pick a single paradigm for all my files.
Heh, funny, looks like I'm both. I make use of searching (spotlight, on windows the search thingy in the start menu) but I also navigate through my project folders when I know where something is (which I usually do with my work related stuff).
I'm also really missing the windows 95/98 start menu that would open sub menus instead of just being a view where stuff slides around.
> Most people I've talked to that experienced windows 8 would agree that when Microsoft did just that, it was a misstep.
Actually Windows 8 can't be a misstep because it led to UI optimization for Windows 10. Windows 8 was just a drastic change for sure, but once you understood it's concepts it was really well suited for beginners and power users alike. The only misstep I can think of with Windows 8 is calling the Start Menu a Start Screen. That made it seem entirely different, when it wasn't.
If windows 8 treated the desktop as the most basic place a user could go in the US and treated metro apps as full-screen versions of desktop apps I think it would have been a brilliant OS.
I'll just reply to this comment since, I basically want to echo the same things. One sure way to get me interested in a Linux variant would be to iOS'ify the Mac Desktop. I can think of few things less interesting or desired from my Mac experience.
Overall, what a disappointing set of ideas put forth by this article...
> Also, removing the file system tree would be a huge step backwards.
Absolutely agree. I also fully believe natural text search is the future for computer interfaces. But what this article ignores is that you can easily have both. I use Spotlight to instantly full-text search over all of my files, applications, and their metadata to launch / open / access my data, but I also keep a standard directory structure instead of keeping all my files in one place. I've started thinking of the directory tree as a hierarchical tagging system, and when I'm trying to find a file for which I can't remember the right search terms it's incredibly useful. No need to throw out the baby with the bathwater; multiple views / interfaces to the same data is very useful.
I suspect Apple was thinking along the same terms when they introduced the "All My Files" shortcut in the Finder. But I'm sure they realized that the majority of their customers (who are barely able to use their computers; more or less computer illiterate) are just not ready to move beyond the "Desktop" / "Filing Cabinet" metaphor.
Actually, I've been thinking about this a lot recently, and I just started using the OS X universal Help menu feature [0] to take actions in complicated applications like Photoshop. Full-text search over all possible menu commands in any application? Yes, please. I really miss being able to do this when I'm in my Debian or Windows environments.
> and the common "solution" with tagging/labels are just single-level folders in disguise....
Tags are best when used to complement the hierarchical-folder-based system; as an extra layer of organization, not as a replacement.
Tags effectively let you keep a single file in multiple "folders", because the reason you're looking for a particular file is not always the same.
For example, my girlfriend is a concept artist and we're both fans of anime, so we trade a lot of reference and "reaction" images back and forth. Without tags, we'd be limited to keeping each file in a single folder, usually the name of the series OR a single character, but we're often looking for images depicting specific emotions, such as "Happy", "Shocked" and so on, and most images usually have multiple characters. With tags I can just save the image under the folder for a particular show, and just tag it with the characters' names and emotions. Keeps everything tidy for the OCD in me and makes searching for anything a snap as well. :)
OSX and iOS need to converge into something new and unified. Touch and pen(cils) is becoming standard for new computers in the near future, and tablets with keyboards also need to support touchpad and mouse. It may not have been necessary earlier, but with Surface and iPad Pro it's becoming obvious that these are not completely separate platforms any longer.
This is way better for e.g. my mother's use case. It's horrible for power users. I would be fine with this if Apple introduced some kind of "power user shell". I kind of do this already (on Win, Mac, and Linux) by almost exclusively using a terminal window and browser.
edit: "macOS" is absolutely brilliant though. They should go for the throwback hipster vibe.
> It's horrible for power users. I would be fine with this if Apple introduced some kind of "power user shell"
That sounds like it could work in theory, but I dunno if I'd actually want that as a power user. Mostly because there's more "normal" users out there, I feel like any alternative UI would be neglected. Seems like you'd be consigning yourself to being a second class citizen.
Maybe this works on Linux where there's less of a profit incentive because development on niche WM's and such is largely done by hobbyists, but I'm having a hard time imagining Apple spending a lot of resources on developing something large for a minority of users. Maybe if they opened it up a bit and had some sort of open plugin based build-your-own experience that could shift some of the effort off of them? I could see that working but it doesn't seem in line with Apple's philosophy.
How would you expect 3D Touch to work on a screen that moves if you press on it? You’d need to maintain the laptop screen in one hand while 3D Touch-ing with the other; that’s cumbersome.
3D Touch/Force Touch actually currently works on OS X on the 2015 MacBook Pro and the new MacBook. The touchpad doesn't move, it's just pressure sensitive and provides haptic feedback so you feel like you can depress it even more after clicking. OS X provides a dictionary/lookup popover when using it in Safari and some other apps.
This is like looking at Porsche try to make a flatbed truck. It's stupid because it belays all the use cases that the product ctaegory needs.
Designer: design isn't just about looking sleek, it's primarily about meeting the use case as simply and robustly as possible. All these use cases are better on other devices.
This is clickbait. There is nothing in here that comes (anywhere) close to XI -- at best, it's OS X 10.12. OS XI will be a paradigm shift -- the way that OS 9 to OS X was -- not just a slightly different UI, desktop picture, and icons.
This fails to provide any motivation for such a merging/converging iOS/MacOS experience.
Why do I want Netflix and CNN apps that have been so easily (but nonetheless with effort) adapted to support the platform from their iOS counterparts? I expect I'll choose the website in Chrome every time.
Personally I’d love to have native apps for media-intensive and “app like” websites. Why? Well, for those sorts of sites a native app will easily perform better and their existence would allow me to dedicate my browser to more lightweight/traditional sites focused on delivering text and images.
1. Fix Finder.
2. Integrate the iDevice simulator into the OSX: all iOS apps can run on OSX. So let them.
3. Continue making kick-ass laptop/personal computers, with keyboards and screens and touchscreen, and make the difference between iOS and OSX go away, at the hardware layer.
Plan B:
1. Put Xcode on iOS.
2. Make an iOS Touchscreen Laptop (8-core ARM with 16gigs RAM, etc.)
3. Abandon OSX completely.
68 comments
[ 3.1 ms ] story [ 97.1 ms ] threadApple will just keep improving iOS to be more powerful, more flexible and so on. The iPad Pro with iOS 9 is just a rough draft, it will be iOS 10 or 11 that'll show the hardware off, at least I hope. iOS 9 was probably the biggest jump in productivity for the iPads and hopefully, Apple keeps it going from now on, not neglecting it like it did prior to iOS 9.
IOS would have to evolve a lot and in many ways to take over desktop. Maybe it will. Android N seems to be moving there so maybe that's the future.
I can't think of a single major thing.
At one point they finally got multi monitor support working properly, but I forget when.
Also it removed some skeuomorphism, which depending on whether you like it or not counts as a major feature ;-)
For developers I think the future is cloud based web dev setups anyway. I really like kicking off a large build or run on a remote server and not have it heat up my laptop. For different purposes, having a few beefed-up VPSs available and services like nitrous is more convenient. That said, developing in Emacs or vi in SSH term windows lacks some convenience, so it having deep integration of IDEs like IntelliJ with your own server build processes would be interesting (and even more interesting to also have great iOS and Android support).
EDIT: more on the topic of OSX: I much prefer the small continuous update process that OSX and Windows 10 seem to be taking. I use Ubunutu and I often think that I would prefer small continuous updates than 6 month drops - at least for laptop use.
Tablets and smartphones have a touch screen, computers most often don't. Software that doesn't take this into account will be a second-class citizen on the respective platform.
And while a flat filesystem with a powerful search function certainly has its uses, I wouldn't want to use it for everything. I'm far too anal when it comes to organizing my source code, for example.
The article touches on a concept called 'loops' for cases such as this.
I think there's room for innovation with filesystems. Personally I'm a fan of the 'database filesystem' ideas from BeOS/Haiku.
How do you see that working well?
Microsoft tried to incorporate all the stuff like MP3 metadata into explorer way back when (XP? Vista? Can't recall)
OS X has supported "saved searches" which act like a live database-backed file manager ever since Spotlight was added. More recently they added quick tagging of files.
What are you imagining that would be more usable than these?
I have my music, my music app queries the database, presents my with my music, tags them into playlists etc, great.
I have my photos. OK, so I want to keep my screenshots and the memes I downloaded from imgur separate, but those are fairly easy to special-case.
Now I'm writing a novel in my spare time. I got my text document, I got whatever Evernote/OneNote-style system for notes. I found this cool MMORPG map editor I'm using to keep track of where my characters live and go to work, does that count as a picture. And I recorded when I hummed a few bars because I wanted to add a song my character wrote when he was bored, and now it keeps shuffling that in my music. Hmm these things are getting difficult to associate now...
And I might be a bit pedantic, but filesystems are databases. Indexing and navigation methods sometimes differ, but a lot of people use advanced indexing in windows that not only checks extended attributes but also contents. etc.
Well all the PC's in my home have touch screen and I surely miss it when I use a Mac at work.
There's absolutely no way you could build an app in a big "bucket 'o' files". Sure, you can put together a flyer for a softball game or (maybe) type up a simplistic report for school, but eventually these "loops" are going to become big, unusable buckets of content that isn't easily searchable (do we have good search on music contents yet? better hope the filename is good before it goes into the bucket!).
Perhaps the answer here is a "subloop". That's starting to look like we've come around to nested document folders again, isn't it?
Somewhat amusing: the screenshot of Sublime Text in the mockup is using folders.
Analogies I'm thinking of include Smalltalk images, or how a compiled .app is a directory right now.
Like the author, I worked on a distributed cloud FS, but hierarchy was a core goal. The abstracted reference system of HFS that split meta-data from the data resources is a huge boon to handling various common issues, from rights management to doing "fast" full-tree cloning over a distributed store and several other neat tricks like persistence. The biggest issue with distributed heirarchial file-systems is probably how to handle the full-tree cloning, in a way that isn't crippling, slow, or causes an enforced depth and file limits.
Organizing files into projects is waved away by "just search through an amorphous blob of data". "Kind: PDF". Who finds files like that?
The user activity thread is nice, but it's clear Apple was already working in this direction with the links timeline in Safari, but all the social service TOS's forbid this kind of presentation and they're super-protective of their presentation to the point even Apple can't negotiate past it at this point.
I get hundreds of emails a day, work in an editor that is indexing ~11000 files, regularly have 50 tabs open in Chrome, etc. I find many apps that follow the sorts of aesthetics and principles shown in these mockups just do not scale to what I would consider normal professional use.
- A "people focused" email experience doesn't work very well with monitoring emails. I've had to turn off Gmail's smart inbox because me being CC'd into a person to person thread that I don't care much about is "high priority", but that Monit email saying CPU is at 99% on one of the webservers looks like every other Monit alert, and is a low priority update.
- I use Slack extensively at work, but email is still incredibly important.
- I may not need 50 tabs open, but after my 5 pinned tabs, a few reference/documentation tabs, the prod version of what I'm working on, the local version of what I'm working on, the PR page, and a few tabs of personal stuff, I'm easily reaching 20+, all it takes is one context switch to a different project and I've opened another 10. 50 isn't uncommon.
- A hierarchical file system is pretty important when working with Git repos. And I'm not a contractor, but if I were, separating client data is of paramount importance.
Also, removing the file system tree would be a huge step backwards. Organizing hundreds or thousands of files between dozens to hundreds of projects is just not suited for a "everything in one searchable bag" for a few reasons: To find something with search, you have to already know what you're looking for - something that browsing a folder solves instantly (how can you know you see all relevant files with a search? Maybe there's an important file that didn't match your query?!) - and the common "solution" with tagging/labels are just single-level folders in disguise....
I'm glad Apple decided to not port OSX to the iPad, or replace OSX with an iPad-esque iOS. Most people I've talked to that experienced windows 8 would agree that when Microsoft did just that, it was a misstep.
Losing a filesystem hierarchy is loony toons. I can't even
That really struck me, because I'm of the latter type, and it makes me somewhat uncomfortable to have a search paradigm thrust upon me.
I can avoid that by having my files in a content-based-addressing pool with metainformation.
I'm not organized by nature, but I need some kind of hierarchy for my computer files. There's too much.
Some collections of my files are accessed entirely hierarchically, others are not, I don't think I should have to pick a single paradigm for all my files.
I'm also really missing the windows 95/98 start menu that would open sub menus instead of just being a view where stuff slides around.
Actually Windows 8 can't be a misstep because it led to UI optimization for Windows 10. Windows 8 was just a drastic change for sure, but once you understood it's concepts it was really well suited for beginners and power users alike. The only misstep I can think of with Windows 8 is calling the Start Menu a Start Screen. That made it seem entirely different, when it wasn't.
Overall, what a disappointing set of ideas put forth by this article...
Absolutely agree. I also fully believe natural text search is the future for computer interfaces. But what this article ignores is that you can easily have both. I use Spotlight to instantly full-text search over all of my files, applications, and their metadata to launch / open / access my data, but I also keep a standard directory structure instead of keeping all my files in one place. I've started thinking of the directory tree as a hierarchical tagging system, and when I'm trying to find a file for which I can't remember the right search terms it's incredibly useful. No need to throw out the baby with the bathwater; multiple views / interfaces to the same data is very useful.
I suspect Apple was thinking along the same terms when they introduced the "All My Files" shortcut in the Finder. But I'm sure they realized that the majority of their customers (who are barely able to use their computers; more or less computer illiterate) are just not ready to move beyond the "Desktop" / "Filing Cabinet" metaphor.
Actually, I've been thinking about this a lot recently, and I just started using the OS X universal Help menu feature [0] to take actions in complicated applications like Photoshop. Full-text search over all possible menu commands in any application? Yes, please. I really miss being able to do this when I'm in my Debian or Windows environments.
[0]: described here http://lifehacker.com/5592856/use-os-xs-application-help-men...
Tags are best when used to complement the hierarchical-folder-based system; as an extra layer of organization, not as a replacement.
Tags effectively let you keep a single file in multiple "folders", because the reason you're looking for a particular file is not always the same.
For example, my girlfriend is a concept artist and we're both fans of anime, so we trade a lot of reference and "reaction" images back and forth. Without tags, we'd be limited to keeping each file in a single folder, usually the name of the series OR a single character, but we're often looking for images depicting specific emotions, such as "Happy", "Shocked" and so on, and most images usually have multiple characters. With tags I can just save the image under the folder for a particular show, and just tag it with the characters' names and emotions. Keeps everything tidy for the OCD in me and makes searching for anything a snap as well. :)
It wrecked Pages, which was one of my favourite software tools prior. I use very little Apple software now.
...Ubuntu here I come.
edit: "macOS" is absolutely brilliant though. They should go for the throwback hipster vibe.
That sounds like it could work in theory, but I dunno if I'd actually want that as a power user. Mostly because there's more "normal" users out there, I feel like any alternative UI would be neglected. Seems like you'd be consigning yourself to being a second class citizen.
Maybe this works on Linux where there's less of a profit incentive because development on niche WM's and such is largely done by hobbyists, but I'm having a hard time imagining Apple spending a lot of resources on developing something large for a minority of users. Maybe if they opened it up a bit and had some sort of open plugin based build-your-own experience that could shift some of the effort off of them? I could see that working but it doesn't seem in line with Apple's philosophy.
Completely superfluous, of course, and they are probably going with macOS, but one can dream of a more sensible casing future.
Designer: design isn't just about looking sleek, it's primarily about meeting the use case as simply and robustly as possible. All these use cases are better on other devices.
Why do I want Netflix and CNN apps that have been so easily (but nonetheless with effort) adapted to support the platform from their iOS counterparts? I expect I'll choose the website in Chrome every time.
1. Fix Finder. 2. Integrate the iDevice simulator into the OSX: all iOS apps can run on OSX. So let them. 3. Continue making kick-ass laptop/personal computers, with keyboards and screens and touchscreen, and make the difference between iOS and OSX go away, at the hardware layer.
Plan B:
1. Put Xcode on iOS. 2. Make an iOS Touchscreen Laptop (8-core ARM with 16gigs RAM, etc.) 3. Abandon OSX completely.