“Intel’s CEO, Paul Otellini, hinted that Microsoft is trying to unify their operating systems into one OS that runs from phone to the desktop, his remark raising questions on Windows and Windows Phone 7’s future.” Obviously, the Windows Phone 7 userland probably runs just as fine on Windows NT as it does on Windows CE, so I honestly don’t get the issue here.
They will unify their platform for development and have Silverlight play a larger roll in both systems (as well as .NET). But they server very different funtions (CE and desktop windows. i mean, CE is real time and bla bla bla). it’s just not in the cards, and that’s a good thing. trust me.
Not only that much Windows CE from what I have heard can *REALLY* scale down to systems with a few kilobytes – I simply don’t see Windows NT ever getting to that sort of capability nor do I think one would ever want that given, as you pointed out, the two are completely different beasts. Regarding the point you raised in reference to Silverlight, there are rumours that the ‘App Store’ Microsoft has planned will have restrict requirements that will require developers to upgrade their code bases or find they’re left out in the cold.
you’re wrong and lucky for microsoft this time, they are not.
I’m really not. While Windows phone platform may be based on the tech that powers Windows 8, windows CE will never(?) be replaced be that tech. You just might not see CE being the core of the mobile phones and slates since they have become powerful enough to handle full blow desktop operating systems (and windows 8 will be very very resource friendly).
And I am rarely ever just flat out wrong
…if they will at least unify the GUI.
Scott Forstall over at Apple is trying to do just that.
Apple basically has one OS, but the layers around it differ depending on where it is. It makes sense to try and have one GUI, but one that adapts depending on the environment. An iPhone for instance has no need for windows, the iPad a little more and certainly the desktop environment does (unless there is a clever way to remove that need). I do note that Lion finally allows Apple apps to go full screen (just like Windows has done forever).
Maybe we are starting to see this in the Linux world with the slow move to QT (I don’t know, I don’t follow Linux as closely).
I would love to see MS have Windows 8 as a unifying OS, one that can adapt depending on the use. I guess going to ARM is a sign of that perhaps. I would love MS to remove all the crap that is still in there from the 95 / 98 days (create a compatibility mode for older apps) and have a fast slim efficient OS. Having one GUI is the icing on the cake.
I do (follow Linux closely), and I don’t see a move to a unified GUI. Indeed, I see just the opposite.
At least 3 major GUIs are based on QT currently – the venerable KDE, Canonical’s new Unity (currently just on netbooks, but coming soon to their full range of consumer products), and Intel / Nokia’s MeeGo. (I’m ignoring the range of non-QT GUIs here for simplicity, since you specified QT.)
MeeGo has a unified set of concepts and theming, but it has distinct GUIs for smartphones, media phones, tablets, netbooks, desktops, TVs, and vehicle information systems. Given that the input devices, screen dynamics, and use cases are very different across this spectrum of devices, I think that makes sense.
So rather than consolidating on a one-size-fits-all GUI, Linux is moving to targeted GUIs built around common control schemes – rather like automobile user interfaces of today (how do I turn on those wipers again?!?).
As to scalability, the Linux kernel runs devices as small as a watch to virtually every supercomputer on the planet. So there’s no conceptual limit on the NT kernel’s scalability. Rather, the questions are whether the NT kernel has inherent design limitations on its scalability, and whether Microsoft can make more profit for themselves with a single kernel than their current rather confusing (to me) mix of technology.
Slow move to QT? what the hell? I don’t use QT much at all and I doubt I ever will. As for consolidation, no way man. QT, GTK, Gnustep all are different and equally valid platforms.
Edited 2011-01-20 11:26 UTC
I’d be happy if they would just offer one version of their desktop OS. Six versions (or however many there are) for Windows 7 is *so* confusing.
My prediction is that the core (Kernel and other absolute necessities) will be merged across all the consumer-facing, server, and higher-end embedded stuff — and mostly the UI will distinguish between them. This is definitely good for consumers and for developers. Now if we can convince them to adopt fewer SKUs, we’d be in good shape — maybe Datacenter/HPC, Server, Business, and Home. Then non-retail SKUs for home servers and media centers.
I predict that they’ll keep CE around for non-user-oriented embedded devices, and shift its focus towards real-time systems. Windows CE certainly scale down much further than NT (and, IIRC, can be made to work without an MMU) so its the right fit for some small purpose-built systems.
… because the WindowsCE variants (including PocketPC, SmartPhone and Windows Phone 7) are also used in long term supported products. Just one example: the handscanners used by UPS, Federal Express, Deutsche Post and many, many others cannot be replaced short term.
pica
All they have done is cast doubt on how committed they are to Windows Phone 7.
I took a look at a WP7 phone recently and was very impressed, it’s an amazing UI in much the same way as the iPhone was in 2007. But the iPhone has stagnated slightly.
It was a very tempting proposition and I considered buying one until I read Ballmers comments about Windows everywhere.
The first question is why? Who cares if its Windows (i.e. an NT core) – the users don’t. They want an experience optimised for the device in question which is exactly what WP7 is.
It’s just to cut costs for MS.
The developers care. If smartphones and tablets use the same platform as PCs and Servers, code can be used on all classes of devices. At the moment writing an application for WinCE and writing an application for WinNT are two different shoes.
Additionally porting existing PC applications to tablets or smartphone becomes easier. Just add a new view and you are done.
As a result more application will be available for tablets and smartphones. Which is a win for the users.
I think unifying on WinNT will be a big win. Not just for Microsoft. It is a win for the users, for the developers and for Microsoft.
pica
Not sure it’s that simple. For example a desktop version of an app might make much bigger use of memory for caching data, which a mobile version would not want to do.
I would not want to build an application that has code branches all over the place to behave differently based on the device type, that would be less than ideal.
Are the API’s of CE and NT not broadly similar? The iPhone might be OS X at the bottom layer but the API calls are different to the Mac. Similar but not identical.
Today Netbooks offer an Intel Atom with 1GB of RAM and a 1024×600 pixel 10″ display. A high end PC might offer an 8-core i7 with 16GB of RAM and a 2560×1600 pixel 30″ display.
If I look at these numbers, I see a factor ~20 in both memory and compute power we have to deal with today.
On the other side todays tablets offer an 1GHz Cortex A9 with 0.5GB of RAM. That’s only a factor ~2 away from a netbook.
Make the amount of resources used for caching, speculative executing, etc. adaptive and you will have less pain in the future.
Yes they are similar — both have API methods for e.g. opening a textbox — but they are not identical. Using .NET the pain is significally less, but even the “normal” .Net framework and the .NET compact framework differ.
pica
Edited 2011-01-19 11:14 UTC