Supersonic Man

December 2, 2012

the end of Windows hegemony?

Filed under: Hobbyism and Nerdry,Rantation and Politicizing — Supersonic Man @ 12:30 pm

Are we finally seeing the first signs of the end of Windows?  Can the vast decaying empire of the Windows desktop finally be about to fall?

Background: Microsoft has always, since its early days, had a bad habit in its software architecture: first, they design something that meets immediate needs but doesn’t expand well into the future, and second, instead of smoothly evolving the now outdated initial approach, they replace it with a new incompatible system, while the old system lives on beside it in a deprecated state, to support older software.  They’ve been doing this ever since the MS-DOS days, when release 2.0 made a whole second API for file access because the first one didn’t support a directory hierarchy, and then several times on the disk filesystem itself as they replaced FAT-12 with FAT-16 with VFAT-32 with NTFS.  Windows has gone through three completely different codebases, underlying roughly similar UIs and APIs — the latter being sometimes compatible and sometimes not.  Individual API sections are full of deprecated functions living on alongside their enhanced successors.

One case that’s relevant here is that DDE was replaced with OLE, which evolved into ActiveX and COM, which was all shoved aside by .NET …the latter being IMHO Microsoft’s finest work, the one time they really did something right, despite it starting out as just an imitation of Java.

Now comes Windows 8.  Microsoft can’t help but notice that PC sales are flat while mobile device sales are booming.  They know they have to get in on that, instead of just being the OS provider for PCs.  Windows 8 (which is actually just 6.2 internally) is their answer to this conundrum.

How does it tackle the problem?  With the new “Metro” style of interface, and the “WinRT” API that underlies it.  Now in some ways, this is a very positive step.  Metro/WinRT has a new approach to API architecture which finally addresses head-on one of the major problems with all previous successful software platforms: the great difficulty of making software that is genuinely responsive, that is always ready to react to the user, no matter how busy the computer is.

In other ways, Windows 8 follows their normal tried-and-true path: let someone else come up with a successful innovation, then imitate it.  The apps that run in this new slick no-waiting branch of Windows are going to be curated through an App Store just like apple’s and google’s.  And they’ll run on tablets, with a new mobile OS called “Windows RT”, and (if designed to) even on phones, with “Windows Phone 8”.  All share the WinRT API, or near variants of it.  The three environments will probably need separate software builds, but they can largely have common source code, I would guess.

So what is WinRT?  Is it, in any sense, Windows?  It is not.  It is an all-new tablet OS, extended downward to phones (replacing the old “Windows CE” codebase that still persisted as “Windows Phone 7”) and upward to the desktop in Windows 8.  Note that I say in Windows 8, not as Windows 8.  Most of Windows 8 is just an incremental update of Windows 7.  The new WinRT part is entirely separate.  The new slick tiley interface for Windows 8 is mostly just an emulator of the tablet OS!  Except with hooks that allow it to launch traditional Windows applications.

What Microsoft has done is to create an entirely new software platform designed for mobile devices, and then make PC sellers include it as a feature with their desktops and laptops.  Software written for this new platform cannot be translated from Windows.  It cannot even really be reasonably ported.  Because of the bold ideas they put in to allow apps to be always responsive and never tied up waiting for some computing or IO operation to conclude, doing a WinRT app properly requires a whole different conception of how to put software together.  Doing it right, when you’re starting from an old PC or Mac program, needs a thorough redesign.  (Though I suppose a lot of people might find ways to cheat on that.)

The weird part of it, technically, is the choice of what previous technology they decided to build WinRT out of.  They already had, in .Net, a highly viable system for maintaining continuity of app development in a post-Windows world, because .Net, like Java, doesn’t have to care what hardware or OS it runs on.  But they did not make WinRT out of .Net — they made it out of COM, the technology that .Net was supposed to have replaced!  I’ve heard rumors that different developer camps inside Microsoft basically didn’t communicate very well, and would end up going in divergent directions and ignoring each other’s innovations.

This COM basis means that developing in .Net needs awkward adaptations, and .Net innovations that could have been valuable, such as the Presentation Foundation, end up disused.  This further undercuts Silverlight, which not too long ago was supposed to be the future of MS phone app development.

Anyway, WinRT might be somewhat successful in the tablet market, maybe even in the phone market.  But nobody seriously expects in either case that it’s going to be the market leader.  Conversely, on the PC just because people who want their Windows up to date get stuck with the tablet environment along with it… but on the desktop, why would anyone want to use it?  The competitor here is Windows itself.  Why would people working on a desk want to pretend they’re on a cramped little tablet?

What Microsoft wants to do later on is come out with a super-tablet that runs full-blown Windows 8.  But such a tablet is likely to be very bulky and have very poor battery life.  And can you imagine how painful it would be to work traditional Windows apps through a touchscreen?  It won’t sell.  So the likely future for WinRT is that it’ll be a more or less viable competitor in phone and tablet space, and maybe somewhat popular in desktop/laptop space too, but it’s pretty unlikely to ever be the big popular hit for either.

So where does that leave Windows?  If WinRT/Metro is a hit on the desktop, then traditional Windows and its apps are now obsolete and semi-officially deprecated.  All the godawful smelly old baggage of Windows programming can finally be discarded, in favor of a fresh modern system — one which is, incidentally, a lot smaller and simpler.  But that fresh new system will now be far less mature and well-supplied with apps than its competitors such as MacOS… or even Android.  There’s no way it can presume to a role of number-one-hood even if it’s installed on a billion PCs, as long as most of them still linger on in being used mainly for traditional Windows apps.

On the other hand, if WinRT is not a hit, then Windows remains the default dominant OS, but it does so from a position of even Microsoft saying it’s now outdated and without much of a future.  People are going to start looking in earnest for alternatives.

What could possibly replace it?  ChromeOS looked like the big threat when they announced it, but so far, it’s a dud.  It’s working out like every other time that some putz has tried to sell a limited, stripped-down computer substitute that can’t do normal computer work.  But that could change with two simple steps: one is to get the price substantially lower than Windows laptops… and that shouldn’t be hard, as I’ve heard it claimed that on some laptops, the Windows license costs as much as the hardware.  The other, which Google has said is an eventual goal… is to merge Android into Chrome.

Android has what Chrome lacks: a rich selection of apps and an API to easily make many more.  Chrome has the one thing that Android lacks for desktop use: windowing.  Chrome without Android is like iOS 1, when the early apps were nothing but specially tuned web pages… an idea that Apple quickly saw they had to move past.  Web apps still just aren’t competitive with real apps.  Android without Chrome, on the other hand, just isn’t suited to a large display with lots of real estate to spread work out in.  Both OSes are small, so combining the two would still be a pretty lightweight system.

Apple could fight back by finding ways to migrate iOS — which still has a commanding lead on the tablet, if not on the phone — into their desktop OS.

The eventual convergence, I expect, will be for traditional desktop computers, and laptops, to make way for desktop stations which are essentially just docks for tablets or phones.  The desktop system will give you the big screen and keyboard and wired connections you need, but your personal files that aren’t in the cloud will stay on the mobile device.  The desktop may supply extra CPU resources but the controlling OS will remain on the phone, because people want to take their desktop layouts with them.  Cloud-based data syncing is not making a lot of headway, and remains somewhat inconvenient.  People who need that now mostly manage it by buying compact laptops and plugging them into external monitors and keyboards.  An optionally-keyboarded tablet is just the next step.  The performance disadvantage of a tablet over a desktop is going to decrease further.  A device with large heatsinks will always be able to outperform a portable one, but there’s no reason why that extra performance will need its own OS.  That extra processing power could even just be “in the cloud”, rather than on your desk.

It’s pretty hard to imagine Windows being predominant in such a future.



  1. Addendum: six months later, it’s looking like Windows 8 is every bit the huge flop many were predicting. It’s sold a lot of licenses, but for business use they’re just installing Windows 7 on machines that have Windows 8 licenses, and for home use, people are almost exclusively using the old Windows part and just treating the Metro tiles as an obstacle to get past. Programs to reintroduce a Start button are highly popular. It is rumored that Microsoft, in the forthcoming Windows 8.1, will be adding back the Start button, and making several other changes based on listening to the negative user feedback that they wilfully ignored before. Adoption of apps from the WinRT app store remains minimal. On the desktop, typical users only open about one RT app per day, and even on laptops with touchscreens, the figure is barely over two per day. Sales of RT-based tablets and phones, small though they are compared to iStuff and droids, may actually be the brightest part of the picture for the RT venture.

    Comment by Supersonic Man — June 20, 2013 @ 4:33 pm | Reply

    • Update: “adding back the start button” turns out to mean only a fake start button that goes back to the metro tiles page.

      Comment by Supersonic Man — September 5, 2013 @ 3:05 pm | Reply

  2. Correction… Windows Phone 8 is making some headway, and people are buying some touch-enabled devices running full Windows 8, but the Windows RT tablet OS appears to be an utter market failure. People are not going to put up with Windows RT in a tablet or notebook unless it can run real Windows apps, which means they need full Windows 8. It looks like it’s only on phones that people are willing to tolerate an RT-like interface without full Windows compatibility.

    Comment by Supersonic Man — July 24, 2013 @ 11:55 am | Reply

  3. A year on, the Surface RT tablet is starting to be quite a bit more successful… but only because of steep price cuts. The Surface Pro, Surface 2 Pro, and the RT Surface 2 are all pretty insignificant sellers, apparently just because they cost more. The cheap option, the first-gen Surface RT, may now be catching up to Samsung’s in-use market share for tablets (still far smaller than Apple’s), but Samsung isn’t sitting still: they’ll reportedly offer a seven inch tablet for only $130 soon.

    Comment by Supersonic Man — January 7, 2014 @ 10:35 am | Reply

  4. Okay, $130 for a Samsung tablet was overoptimistic.

    Windows Phone is being taken seriously, and I see people use them, but the latest figures I’ve seen say that it’s only in the last few months that they’ve passed the US market share of the dying Blackberry.

    In tablet land, several companies are now selling WinRT and Windows 8 tabs. Microsoft’s share remains dinky, like 2.5%, but is expected to grow as ultrabooks become increasingly tabletlike, legitimizing things like the Surface as alternatives. Other sellers of Windows tabs might add 2% more. Microsoft is rumored to be weighing the option of putting a subsystem in Windows to run Android apps. Meanwhile, they’ve announced the Surface Pro 3, a twelve inch tablet designed to be powerful enough to actually replace a laptop.

    Up at the front of the pack, Samsung is closing in on passing Apple’s market share; Apple is losing ground there more rapidly than in phones. The whole tablet market is being hurt by the large size of current phones.

    Comment by Supersonic Man — May 7, 2014 @ 5:31 pm | Reply

  5. I have still never used Windows 8, but I now have Office 2013. The W8-inspired design decisions they made for it are just awful. They’ve really increased the difficulty of seeing what you’re doing.

    Comment by Supersonic Man — May 16, 2014 @ 9:59 am | Reply

  6. At first it looked like RT might be doing better than the Surface Pro type of tablet that runs full Windows 8, but that’s now reversed. It’s full Windows 8 that’s having some mild success in tablet-land, while RT languishes with hardly any new devices adopting it. Windows Phone 8 may be gradually clawing its way from a 3.5% market share to a 7% market share.

    The coming Windows 9 may be less hated on the desktop than 8, but it doesn’t sound like it really moves forward much. As tablets become more powerful we may see a gain in smaller and lighter ones running full Windows. Microsoft is trying to encourage this with price breaks.

    Comment by Supersonic Man — August 21, 2014 @ 10:31 am | Reply

  7. The market share of Windows Phone apparently peaked at 3.3%, and has since dropped to 2.7%. They’re doing worse than almost anyone dared to expect. The tablet share for Windows is healthier at 4.6%, and might do a lot better if it weren’t for the negative public opinion of Windows 8. So Windows 10 (they’re skipping 9) might really help. But tablets in general have slowed way down in sales growth, caught between phat phones, dropping laptop prices, and greater-than-expected longevity for their own earlier models. The hybrid laptop-tablet or “2-in-1” could take off — so far it’s pinned down by hatred of Windows 8. (And they mostly haven’t got the weight below three pounds yet.) Meanwhile, schools are buying cheap Chromebooks now instead of iPads. Chromebook sales are now about the same magnitude as Surface sales, but apparently growing a lot faster. Chromebooks under $200 (or two pounds) are still not ready, but are threatening to happen.

    Comment by Supersonic Man — January 6, 2015 @ 8:11 pm | Reply

    • Windows 10 is being designed to run on either PCs or tablets, with a smaller version to run on phones. In announcing this, MS is essentially admitting that they’re killing the RT-only OS. They say they’ll come up with some sort of version for Surface 1 and 2 users to upgrade to. All future Windows 10 tablets will have to be powerful enough to run legacy Win32 apps.

      Comment by Supersonic Man — February 22, 2015 @ 12:19 am | Reply

  8. I saw a report that for at least a little while last fall, the adoption level of Windows 8.x actually dropped relative to Windows 7. Recent and accurate figures are hard to come by, but it looks like 7 still holds a nearly 60% share. The best figures for 8.x show it just now pulling ahead of XP at somewhere below 15%; other reports say it has a ways to go yet to catch XP. Meanwhile, PC sales overall are no longer flat — they’re now dropping.

    Comment by Supersonic Man — April 26, 2015 @ 8:47 pm | Reply

  9. Huh, the Windows Phone effort collapsed a lot more abruptly than anyone expected. They just flat gave up on competing with iAndrOS on phones, and announced about 7000 layoffs stemming therefrom. So they’ll make Surface tablets, and for phones they’ll hack up a version of Android.

    Comment by Supersonic Man — July 8, 2015 @ 10:45 pm | Reply

    • Meanwhile, 13″ chromebooks are touching $200, and 11″ers are around $160.

      Comment by Supersonic Man — July 9, 2015 @ 2:10 pm | Reply

  10. Windows Phone is still being sold, but the market share is below 2% now and it’ll probably keep fading fast.

    On PCs, Windows 10 has now passed Windows 8, but is still a long way from catching up to 7. But with all those combined, Android devices are now outselling Windows devices by a ratio approaching four to one.

    ChromeOS is not catching on at all — it remains below 1% in browser stats, which is exactly where it should be strongest.

    Comment by Supersonic Man — March 18, 2016 @ 1:23 pm | Reply

  11. ChromeOS may not be big in the market yet, but it’s apparently becoming a default experience for young students, which bodes well for its future acceptance. ChromeOS has climbed to about a 2% share in the USA, though it remains much lower worldwide.

    Meanwhile, Windows 10 has almost caught up to Windows 7 worldwide, both having four times the share of Windows 8.1, and about ten times that of lingering XP installations. In the USA Windows 10 has passed 7 and 8 combined.

    Comment by Supersonic Man — November 1, 2017 @ 6:57 am | Reply

  12. Microsoft says that Windows 10 is going to be the final Windows — perpetually self-updating, so there will never be another major version release. Kind of a shame because it’s a pretty sucky place to leave the once distinguished OS. A lot of stupid and faddy design choices (don’t get me started on how dumb and overused “flat design” is) are now going to be difficult to move on from. A real competitor to Windows is still nowhere to be seen, but it’s clear now that even Microsoft is just going to let it coast rather than trying to zoom forward with it.

    Comment by Supersonic Man — December 29, 2017 @ 11:18 am | Reply

RSS feed for comments on this post. TrackBack URI

Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

Create a free website or blog at

%d bloggers like this: