Is WinRT going to be the go-to means for all application needs ? from interactive apps to system maintenance and "server" applications ?
Is the intent that you should be planning to change over to WinRT even for services (webservers, ftp server, database server, etc) and typical maintenence apps (anti virus, disk defragmenting, backup/restore, ...) somewhere in the forseeable future, or will those continue to be built around the Win32/Win64 api's with WinRT for interactive apps only ?
Digging around a bit more I finally did find the answer. Which is:
WinRT/Metro is for interactive apps only, services will continue to be the realm of Win32/Win64.
I would call this a prediction (rather than the answer).
WinRT is introduced as an alternative to Win32. It can be considered for any application category. If it fits the bill you're free to use it. This means only the future can tell how and for what it's going to be used. And that's seldom what the prophets envisioned.
WinRT applications that aren't active are NOT getting any CPU time other than what they need to update their tile. This alone makes them unsuitable for services or even certain types of apps that do their thing in the background (backup, defrag, virus scan, ...)
Add to that the fact that each app is sandboxed, and the lack of ability to communicate to the rest of the machine (well for the most part at least) shuts it out for any other type of service.
At least that's what I make out of it, I'd love to hear from someone that has done actual winrt/metro development and got his app published to tell me otherwise.
Windows Store apps are indeed running in a sandbox, so writing system applications with them will be very hard and sometimes impossible.
Running Windows Store apps in the background is also not always possible, only some types of apps are allowed to do that, such as media players.
I'm 100% sure that the desktop part of Windows will never disappear. The Windows Store apps and normal Windows Desktop applications will always coexists.
The reason is obvious, some applications simply cannot be transformed into Windows Store apps. As said system applications are one example, but also complex desktop applications. Remember, Windows Store apps are all about simplicity, easy to use. Look at applications like Visual Studio, Photoshop, Premiere etc etc. I'm sure you can develop a stripped down version of these applications as Windows Store apps, but for sure it's not possible and not recommended to put the complete functionality of those products into a Windows Store app.
Well, currently you cannot publish a WinRT app in the AppStore if it breaks out of the sandbox, that's true. But that's just a policy limitation, not a technical one.
Money is pragmatic. As soon as Microsoft realises there's more money to be made if they lift this artificial restriction they'll do it in the blink of an eye. I wouldn't be surprised if you can sell any application in the AppStore in the not too distant future; Even pure Win32 ones that neither are sandboxed nor even look like a Metro app.
Only the future can tell how WinRT will be used, if at all. But one thing is for sure, most predictions are doomed to be wrong. My favourite is the one made by Ken Olsen. He was prominent CEO of DEC, one of the dominant and most admired computer companies back in the seventies. In 1977 he predicted in response to the increasing popularity of PC's that "there is no reason for any individual to have a computer in his home". Needless to say, DEC is now long gone and forgotten.
Modernity is an elusive target. It cannot be demanded. Metro is new, that's all it is, and only for now. I think it was a big mistake to marry WinRT with a specific UI. It will be a painful process to divorce them but the sooner they go their separate ways to live happily ever after the better. Now that's modern.
Last edited by nuzzle; October 8th, 2012 at 03:39 AM.