Why a merged Apple OS is one mash-up too far It's come full circle. Young reporters joining a prestigious tech publication 20 years ago were quietly advised to focus on only three companies in what we then called "Client Computing". Who were they? Intel, because no other chip company mattered – nobody could produce powerful PC chips at scale. Microsoft, because, well, it …
Having a common kernel is a good idea, what Apple should do is think about different personae depending on form factor which will have to be adhered to my software vendors, I.e. a cut down interface with different functionality in mobile mode and a full desktop experience when linked to a monitor and keyboard/mouse.
I actually quite like some of the touch features in windows 10 in desktop mode
Nothing wrong with that nope... it's the merged GUI that is bad!
In truth, behind the hood can be different at times. A GPU for science/gaming etc could have a really complex pipeline that a Phone may not need/have... but in principle I don't see that stopping the code/design path being the same. As long as it "scales" correctly.
"As long as it 'scales' correctly"
Unlike 'Ape' and Win-10-nic, which are/were "phone OS on a desktop". BOOOOooo!
from article: "Most computing isn't done on a PC any more"
This depends on how you define 'computing'. I do not think that word means what you think it means.
I guess I'm 'ok' with it as long as Apple doesn't cave in to the 'Google Chrome' and 'The Metro' 2D 'FEELING' FLATSO. If they _DO_, it will be the *SINGLE* *BIGGEST* *MISTAKE* *THEY* *COULD* *EVAR* *MAKE*!!!
(Jobs would NEVER do that, if he were still here, as demonstrated by what he DID do, which was VERY successful!)
ok some apple things look a bit flatter than they USED to, but it's STILL at least SOMEWHAT 3D skeuomorphic, but if they go full-on Larson-Green or Australis UI on us, their fate will be SEALED.
Then again, maybe not. Apple fans will probably have it shoved at them until they "like" it, similar to how RIAA markets their "music" these days, and how hollywood continues to pump out CRAP instead of movies (and then give one another mega-awards for making crap movies that nobody watches).
"Then again, maybe not. Apple fans will probably have it shoved at them until they "like" it, similar to how RIAA markets their "music" these days, and how hollywood continues to pump out CRAP instead of movies (and then give one another mega-awards for making crap movies that nobody watches)"
I don't think they will take it. When the 2016 range of MacBooks was launched, there was a lot of blowback from professional users about the loss of useful ports, lack of upgradeability, outdated chipsets and the rest. Judging from various Mac forums, that was too much for many Mac users and they defected to other operating systems.
The danger for Apple there is that once Mac power users leave the Apple ecosystem, they take their families with them and Apple loses them for good, i.e. no need for iTunes or iPhones any more. Apple ought to tread very carefully now or risk losing more customers.
"In truth, behind the hood can be different at times. A GPU for science/gaming etc could have a really complex pipeline that a Phone may not need/have... "
The core Darwin kernel always has been the same between Apple products. I imagine the only real kernel differences are hardware drivers and perhaps default process priority settings. The overall OS differences are the GUI (obviously), filesystem layout and pre-loaded programs ... sorry, "apps".
Those with experience with Linux will quite rightly point out that a different window manager can completely change how a device works.
You need a bit more intelligence to get the application to handle having two faces (if appropriate) but having One OS and two window managers doesn’t seem like an inherently bad idea.
"You need a bit more intelligence to get the application to handle having two faces (if appropriate) but having One OS and two window managers doesn’t seem like an inherently bad idea"
I agree with that but if there's the same interface for the mobile and MacBook then that's probably not going to go down well with users who have to do proper professional work.
There are already alternatives already out there like Windows 10 for the creatives and Linux for the developers and they will surely gain if Apple doesn't handle this one competently. I would have trusted Steve Jobs to pull this one off but with the current leadership crew not so much.
Linux runs Android and can run desktops. So a common kernel is already a reality; we just need to argue about how far up the stack up we want to go. And on the web, a single site can adapt between desktop and touch. (To what extant and how well is in the lap of the designers).
Tablet computing, itself, is a precedent for Apple perfecting what Microsoft has screwed up.
And there is an unmet need for portable, personal computing: I just like the idea of being to plug a device into a keyboard and mouse (and perhaps a monitor) and to have a desktop OS that's mine. So perhaps the time is ripe for Apple to come along and "invent" this.
I am afraid that in this case, the convergence of the OS also means a convergence of the UX. In particular, I'm afraid that because most Mac users use it for very little serious work, the UX is going to converge to a consumer experience similar to the phones.
There has already been a few changes in that direction. For instance, they had in Finder windows this "All my files" folder, which is so useless to anybody doing serious work on a computer that's it's almost insulting. I think that's been removed now (and of course you have people complaining since it was so useful to them), but it's a pretty good indication of what can go wrong when you design for your average users.
A bit like when Windows introduced the ribbon, and the most prominent buttons were "copy" and "paste".
Isn't this where Microsoft was going with Continuum on Windows Phone? Plug it in and you get a Windows PC. Though I'm not sure if it ever offered an actual desktop. Anyone with a Windows Phone who used it able to tell me more?
Continuum gives you a standard Windows desktop.
The start menu just contains your regular home screen tiles, with All Apps off to the right as usual.
Applications are restricted to "fullscreen" only (apparently there's an app that allows you to tile windows across the display), and their UIs reflow to fit the display. There's still a task bar at the bottom though, so it's easy to switch between running apps as usual, and alt-tab works as you would expect.
The Continuum dock (it can work wirelessly too, apparently) has a couple of USB ports and a few video outs too (Display Port and HDMI, I think - I've only used it with HDMI), so plugging in a mouse and keyboard basically works as you would expect. You can still use the phone display as a touch pad and keyboard if you don't have those handy though.
There are a few problems with it - I haven't found an easy way to access things which are wired to physical buttons, like the volume controls, for example, and you can't right click on a task bar button to close an application (you have to activate it, and then click the close button in the title bar), but I think CShell will probably help solve the latter.
Apart from those minor problems and the lack of tiled applications, it's actually pretty fucking awesome IMO. When Andromeda comes out and has Win32 on ARM support, a lot of people really won't need Windows desktop machines anymore.
And that's utterly insufficient. Even with a Surface Pro and UWP applications, same limitations.
When you connect it to a large monitor, maybe more than one, you understand how stupid the idea is of applications designed for single small screens and touch input just "reflowing" to fill the much larger available space, without actually using it efficiently.
You need a far different UI which really takes advantage of the available space and input controls, and widgets may have to change to adapt to the way they're going to be used and how much data can be displayed.
You misunderstood me. All the applications I've tried do have different layouts depending on the aspect ratio and resolution of the display they're running on. It might not be anything other than portrait Vs landscape with components positioned relative to the edges of the display and scaled according to DPI, but that's pretty much all you need most of the time.
Ironically, the desktop monitor (1900x1200) and TV (1920*1080) I have used continuum on have lower resolutions than the actual phone display (1440x2560) anyway.
Linux runs Android and can run desktops. So a common kernel is already a reality…
That's a bit sweeping and, as a result, misleading. The monolithic Linux kernel does support lots of different archs but tends to get customised for each one as a result. Then there is the boon and curse of the different layers, especially when it comes to GUIs which is why both Android and MacOS (some fanboi has just edited one of my questions on AskDifferent to use the "official" writing) mandate their own tightly coupled app frameworks and why lots of people moan about Android not being a full Linux.
With the Mach kernel Apple should have it easier porting the majority of the OS services to different archs and it has been doing for years on simplifying the GUI toolset. Though it probably wants to refactor some of the IOS settings that have grown all kinds of warts in the last few years.
But a toolkit that works for both a touch and mouse interface is notoriously difficult to get right. As always the devil is in the detail of the kind of widgets you want in which environment. I'm seeing this with the Gemini, which is a lovely device (sound on mine could be unusually poor), that is at the limits of a keyboard and touchscreen interface (you can reach the lower part of the screen with your fingers while typing) where a lot of widgets are being caught out on the half-height screen. But you can do a lot and for the rest you can let developers provide different settings information for different widgets for different resolutions. Developers will appreciate this if it is done correctly and one of the reasons for the popularity of QT's QML, I believe.
As for converged devices, I'll believe them when I see them. Apple might well let others test the water and make a splash when it thinks there is a market to be had (it stopped pioneering years ago).
(OS ≠ UX and OS ≠ CPU to put this programming gobbledegook in nicer language.)
I agree, the article seems to confuse orthogonal (independent) issues. The chip that is running a Mac (or iOS device) is orthogonal to the CPU.
A CPU designed from the ground up for today's environment should include security as low as possible to do bounds checking, prevent injected code, etc. (This is the semantic disservice that C has done the industry, separate to the syntactic mess above!)
However, the article makes an important point that mixing touch screen and keyboard/mouse interfaces is like mixing oil and water.
Perhaps the lesson is that most of the industry throws in features without considering what the implications of the mix is, whereas Apple seems to consider those questions before just rushing in.
This happens at all levels - just look at programming in C++, the most awful mixture of everything that anyone can think of, accompanied with many misleading excuses about how powerful that is and that C/C++ do not get in the way of what the programmer wants to do. Actually the mixing oil and water very much gets in the way.
I should also say that Linux running end-user computing is not a good idea. It was a rushed job by Google to use Linux, just because it was there.
However, Linux monolithic kernel architecture makes it less secure. But it is faster. That is OK in data centres where security is monitored by professionals who install software very judiciously. But this manual approach is not good for end-users (even when we're professionals who know about security). We want security to be automatic and to guard against malicious software that we might have - for good intentions - installed.
If it's done right, I could see it working.
It would be nice to have a single computing device that has a touch-screen and fits into a pocket, with a phone UI. Yet if attached to a monitor, mouse and keyboard it has a desktop UI.
Done right, it would need little or no change to application code. Done right, the contextual mode handling would be in OS (well, GUI layer really, but it's all lumped as part of the OS these days) and handled transparently.
Simplified example. App needs a menu bar. It calls the "construct a menu bar" library and populates it. If the device is in desktop mode that results in a standard menu bar being rendered. If the device is in phone mode it gives a hamburger menu button.
Yes, there's a hell of a lot more to it than just that simplified example. A lot of work. But it's not inscrutably complex. Feasible. If you wanted to produce a modal GUI that worked well in both modes, it could be done.
It goes without saying that Win 8 and Gnome 3 did it exactly wrong. You're going to have a common OS and it is going to look like a phone. Even on a desktop. That worked out soooooo well.
We already have realworld examples of devices that change their layout when you change control schemes; what happens when you turn your mobe 90 degrees? And I have a playstation 3 controller plugged into my win10 box, and the computer will accept inputs from it when I open a game with controller support (some games even display keyboard button prompts, but switch to proper controller button prompts when I press a button on the controller).
It seems like we have everything needed to make a proper OS that can switch between UI's for different input setups, we just need someone to do it.
Windows 10 being "awesome" is still a matter of opinion. On the whole I like it (love it, compared to 8), but there's still plenty to hate. The Start menu is still awful, so now I never even open it. I could do at home, where I have Start10 installed, but I've just got out of the habit. Also the new settings "app" is terrible; it's a massive pain to find the right options page to do anything. Thank God the old control panel is still buried in there. Cortana is utterly pointless, and is killed with RegEdit on any PC I work with. The notifications page might as well not exist, since I never look at it. Basically any of the bits that Microsoft are adamant should work the same on touchscreens and with a mouse are total rubbish.
Apple will be fine merging iOS and OSX, so long as the touchscreen mode looks like iOS does now and the desktop mode looks exactly like OSX.
The awesome comment was directed more to the fact that I finally have a device that I can carry around in my pocket that effectively turns into a desktop computer when I plug it into a dock. Yes, there's room for improvement, but it's still pretty awesome. The only thing that could make it more awesome would be if my AI-piloted flying car had a dock and screen. And a bottomless minibar.
I still don't understand the Start menu hate. I suspect a lot of it is just people trying to get their views validated (it's still an easy way to the upvotes around here). Let's compare the start menu in 7 and 10 for a minute...
7: The most recent apps appear in a list directly above the start button - a nice idea, but the most used apps appear furthest away from the start button. (Unless you have your taskbar on the top of the screen, I guess.) Furthermore, the order will potentially switch around if several apps are used as often as each other - reducing the ability to use muscle memory to open applications (albeit ever so slightly). You can remove applications from this list, but you can't explicitly add something there (IIRC). Below all that is the Windows 95 style gargantuan All Programs tree of crap for all registered applications along with all the bullshit web and redundant uninstall shortcuts that the developer (or marketing department) deemed was absolutely necessary for you to run their software.
10: The left half is an alphabetical ordered list of installed applications. (There's still some crap in there because old developer habits die hard, but it's nowhere near as messy as the All Programs menu from before.) The right half is a fully customisable space for your favourite and most used applications. After five minutes of setting this up, you can develop and tone your muscle memory to perfection. (You can even add a shortcut to the old Control Panel there.) I don't remember how the search worked in 7, but I rarely have trouble finding an application by typing a few letters after opening the start menu in 10 - I don't even have to click anywhere first.
Most personal assistants are still pretty useless. Cortana, Siri, whatever the fuck Google has... I don't use them. I can see the point, but I just don't think the technology is there yet. There's a whole host of privacy issues that need solving from a legal perspective before I would consider using them too.
I don't think it's any harder to find something in the Settings than it was in the Control Panel of 7. Certainly if you didn't know where something was in the Control Panel, it could take ages to find something. At least there's a semi-intelligent search box in Settings. The biggest problem with Settings is that it's simply not finished. They should have completely replaced Control Panel in one go - but I realise that wouldn't have been possible with backwards compatibility requirements. On the whole though, I like the fact that settings are now freed from the tiny little cramped dialog boxes of yesteryear. If they can just finish the rest off and get rid of all the bloated peripheral panels with bullshit custom UIs (hello, Intel and Logitech), I'll be happy.
Anyway, not going to try and convince anyone they're right or wrong here, just fancied sharing some thoughts. :) (First <icon>'s on me.)
I dabbled in a Cube iWork10 tablet with a keyboard attachment.
Ran Android and Windows 10. In Windowsland, with the keyboard attached it acted like a little laptop.
Without the keyboard it went into "tablet mode" which mostly got rid of the desktop. Also had a buggy onscreen keyboard that didn't appear most of the time,.
"Of course experience suggests that when a bug is created once, it is propagated everywhere. Apple has been a real master of this lately"
Quality control has gone down in the past few years especially in respect of macOS releases. The late Steve Jobs would not have tolerated that sloppiness.
You could port OSX and all the applications that run on it to ARM in such a way that an end-user would not notice anything different, other than in speed and battery life.
Given that many software suppliers take a different UI design approach for iPhone and iPad, even though they are the exact same operating system, there is no reason why they couldn't continue to also take a different approach for iPad / Desktop+Laptop.
It's not as big a deal with such an OS to change the top layer, the UI,
given the difficulties that Canonical trying to make a new Top layer/UI had I tend to thing that it is a bit harder than most people think it is.
As for apple, their two UI's seem to work most of the time. Moving away from Intel is probably no big deal for them given the ARMiness of IOS.
The problem is not changing the display system and the desktop manager... depending on if it's installed on a desktop or a phone - that's a static problem easily enough to solve. Even Windows is less "monolithic" than people think, although it's evidently more "integrated".
The problem is to have applications that can dynamically adapt to the "environment" they are run in any given moment, maybe even when it switch while in use because you dock it to a keyboard, mouse and larger screen.
And it's not enough to simply display the same application on a larger monitor, just enlarging and re-positioning some chubby elements, you may want a different UI because of the different input methods - you may want to use more complex features and take advantage of the increased precision, for example.
Think about a photo editing application. While in mobile/tablet mode you may want relatively basic controls, especially as long as your input devices is your finger, and a UI usable on a smaller screen.
As soon as a pen, mouse, keyboard and larger monitor is detected, you may want more and more powerful editing tools, and precision ones as well, with more granular control on inputs, and better use of the available screen space.
"given the difficulties that Canonical trying to make a new Top layer/UI had"
As far as I could make out from Unity they were trying to make the same sort of one-size-fits-all Frankeninterface as W8.
OTOH it's perfectly possible to have a choice of KDE/Gnome/Mate/XFCE/whatever at login time. Not, not install time or boot time, login. That's interchangeability.
"As far as I could make out from Unity they were trying to make the same sort of one-size-fits-all Frankeninterface as W8."
And by the sounds of it, this is exactly what Apple are contemplating, one OS from the iPhone to the Mac Pro.
I won't say that they're going to mess it up though. This is Apple we're talking about, they don't always seem to obey the same 'laws' as other tech companies. After all, Microsoft spent years trying to make a tablet PC that people wanted to buy, but it wasn't until the iPad that people actually wanted one.
“OTOH it's perfectly possible to have a choice of KDE/Gnome/Mate/XFCE/whatever at login time. Not, not install time or boot time, login. That's interchangeability.”
I used to use ion3 (tiling WM) as my main window manager, and fire up xnest to have xfce embedded in one of the tiles for the odd occasion when I needed it...
Don’t even need to wait for login (although programs were started in one or the other)
given the difficulties that Canonical trying to make a new Top layer/UI had I tend to thing that it is a bit harder than most people think it is.
Canonical only had difficulty with making the new UI because, like Windows 8, Windows 10, and GNOME 3, it was unsure what it wanted to be, and somehow that always seems to play out as desktop PC users being asked to adapt to a completely inappropriate and needlessly klunky phone UI on devices that, in terms of user input, have little similarity with phones.
Had the Canonical devs followed the Unix philosophy of "do one thing and do it well," they would not have been trying to create one UI for all devices. Hell, if the Unix philosophy is too highbrow for them, try Curly's fireside chat from the 1991 movie City Slickers, in the scene where Curly (played by the legendary Jack Palance) was explaining his philosophy to Mitch (Billy Crystal). One UI to rule them all is one of those ideas that seemed like a good idea at the time, but once attempted, soon revealed itself to be unworkable.
In my diatribes against Windows 10, I've often cited Apple as having gotten this one right by opting out of the unified UI, and I've remarked that Apple would never have released a disjointed mess like Windows 8 or 10. As much as I have always seen Apple as the "bad guy," one thing I cannot fault them for is demanding excellence in their UI design, whatever the particulars of that design may be. If they stick to that, this "half and half" idea will necessarily be rejected once again by Apple as it has been in the past, for the very same reasons Tim Cook and others have articulated.
"given the difficulties that Canonical trying to make a new Top layer/UI had I tend to thing that it is a bit harder than most people think it is."
That's a problem with Canonical's approach, not with the underlying ability to do it. Right now, you have your choice of at least a half-dozen very different desktop UIs for Linux, so clearly it is far from an insurmountable problem.
>>It's not as big a deal with such an OS to change the top layer, the UI,
>given the difficulties that Canonical trying to make a new Top layer/UI had I tend to thing that it is a bit harder than most people think it is.
It wasn't that changing a layer in a well-architected system was difficult. It was that the new design chosen for the layer was 'sub-optimal'.
given the difficulties that Canonical trying to make a new Top layer/UI had I tend to thing that it is a bit harder than most people think it is.
I'm not so sure Canonical underestimated the task so much, but they certainly overestimated their capability - Unity 8 & get phone partners & Mir display 'server' & Snap & other investments....+ day to day support of a Linux distro fairly popular on the corporate stage.
Seems Apple have no idea what to do with an iMac/MacBook apart from to turn it into another iDevice. So far, that means jettisoning features and bashing square pegs into round holes. XServe gone, servers gone, routers gone, time capsule gone, displays gone, crapper UI, server management about to go.
What used to be an ecosystem will probably stagger on until XCode can be made to run on tablets.
the technology was all there and readily available.
Yet they didn't... There was obviously money to be made... errrr...? Perhaps not. Sales of the Microsoft Surface and other Windows Touchy/feely devices have not been exactly stellar now have they.
Apple has been hampered in recent years by the failure of Intel to get low power CPU's out when they say they would. Taking the CPU design in-house when their version of ARM CPU's in iDevices already outperform the competition makes sense when you look at Intel and the limitations of the CPU's that Intel make that are suitable for Mac's. 16Gb of ram does not cut in this day and age.
Taking control of the CPU design is the next logical step for them. After all, they are designing their own GPU's these days.
And there is one less link in the supply chain for Ming Cho to spread fluff about. That can't be bad.
Anyway, where is he? He's been silent on this whole move Mac's to ARM thing. Perhaps there really is nothing going on in the supply chain to back up the posts by Bloomberg who in recent times do seem to be rather down on Apple (like the rest of Wall St).
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
