Should we be starting to develop all desktop apps for windows 8?

[JoePage]

Sorry if you have seen this question too many times before, but with the release of windows 8 it really changes everything. Personally I am not a fan, I think the idea is great for their tablets/phones but for desktop computers it really isn't making much sense to me. I am just unsure where to go right now. I know the future is hard to see right now, but what is everyones take. Have you guys switched over your hobby/commercial projects to windows 8? Will you be developing mainly for that now? I wish I could see the future and how w8 is going to end up in the desktop world to clear some of the fog. I just feel like this is ruining the desktop world. It seems like the only future with windows desktop dev is going to be webbased apps.

[jmcilhinney]

The sky is not falling chicken little. Windows 8 does NOT change everything. Windows 8 provides two interfaces: the phone/tablet style UI previously known as Metro and the desktop interface that is pretty much identical to Windows 7. If you want your app to run on Windows RT and/or Windows Phone then yes, you will have to create a Metro UI for it. Otherwise, just keep doing what you've been doing and your app will run on the Windows 8 desktop just like it ran on the Windows 7 desktop.

I've been using Windows 8 on my desktop machine for several weeks now and just installed it on my laptop last night. I've also had Windows Server 2012 running on my home server for the last couple of weeks, which also has the twin interfaces. I intend to get a Nokia Lumia with Windows Phone 8 at some point in the not too distant future, as well as a Windows RT tablet and a Windows 8 Pro tablet which may be Microsoft Surface or maybe not. I have no issue with the split interface. I'll have to use Metro on Phone and RT and I reckon I'll probably use it primarily when using the Pro tablet with touch too. When working with a keyboard and mouse though, most things I still do using the desktop interface. Some things feel more right on Metro and others on the desktop, so take your pick. Once you get used to having both, switching between them is a no-brainer and does not feel disjointed as some would have you believe.

That's my experience anyway. I'm no Microsoft fanboy but neither am I on the lookout for any little thing to bash them with. I tried it. I like it. I don't feel like, on the desktop at least, that Metro provides a compelling reason to upgrade but it's also not a reason not too. Windows 8 does provide various other improvements though, and those are the primary reason I upgraded. It replaces Virtual PC with Hyper-V and that alone was enough to get me to upgrade.

[dilettante]

Consider that "Windows 8" appears to mean at least 3 things from a programming perspective:

• A new point-release of Windows 6.x with some improvements and changes including the Start Menu factor.
• A new subsystem layered over both Win 6.x and the WinRT-only ARM based Win6.x, almost entirely for low-powered tablets.
• A new "Phone 8" that is closer to the WinRT platform than prior Microsoft entertainment phones (Phone 7).

For many programmers only the first case is where they'll have any real "play" as a result of Windows 8. Unless you shift to the new Microsoft-only tools and the Metro applet programming paradigm you won't be running anything there. But they consider this a "read only" space for almost every user except applet developers anyway. It is very hard to deploy Metro applets onto other machines except through Microsoft's "store" unless you are an enterprise customer.

So ignoring the "pros" (of which there are many in Win8) and the "cons" (already belabored at length and a web search will turn up quite a long list) there isn't much impact on a desktop developer. An exception may be subtle changes to the security model or desktop runtime environment that have not been elaborated on by Microsoft but do seem to prevent some legacy software from installing and running correctly even under appcompat. No, most of the headaches are users' usability headaches.

If history means anything you can ignore the Phone 8 platform which relatively little has been talked about. It isn't clear to me how much Phone 8 development is like WinRT Metro development or where it differs. Phone 7 sales tanked compared to the more popular Windows Mobile product line that preceeded it. I'm not sure why we should hope Phone 8 even reaches the tiny market share (and shrinking) Phone 7 has today.

The ARM tablets (e.g. "Surface") are another question. Whether there is any market penetration remains in doubt. They are too expensive as an impulse buy compared to something like a $100 or $250 Android tablet and have an uphill battle against the pricey but well known iPad.


There are alternatives.

I've taken some steps into Android development instead. The cool thing there is it covers a growing market of phones, tablets, "net tops," set-top boxes, and embedded systems. Though it isn't clear what will occur with Google also offering Chrome OS machines, but Android could easily develop into a laptop and desktop OS - and indeed you can build an Android desktop now using either x86 or ARM.

BlueStacks is a product in early development stages. It alows Android apps to run on Windows, including Metro integration on Metro machines. Right now ARM-emulation makes it slow on x86 machines like your Win7 desktop today. But on ARM it should run ARM code more quickly, and there already is an Intel x86-Android layer for Windows on x86 (HAXM) that may be incorporated. I suspect AMD will do better with its Android-virtualization layer in the future too (this is what is in BlueStacks for x86 right now - only runs well on AMD CPUs).

So if things fall right an Android developer might have a very wide range of platform targets to choose from.

[SJWhiteley]

The only issues I've had with (windows 8 desktop) is that some programs look for a specific OS, either in the installer or in the application itself. Those, sometimes, fail - not because of an actual error, but because the programmer did not accommodate an OS upgrade.

I don't like the interface; the pinks, the purples and the harsh glyphs and the rather unpolished look. But it works, and as a developer, it isn't a real game changer with respect to getting programs running. Programming for the Metro interface is a little different, but it is really a case of extending ones skills, rather than replacing them.

I will say that mobile development is definitely in almost everyones future: the fact that you can use existing skills (VS and .NET) to translate over with minimal effort is quite handy - moving over to Objective-C on the Mac, for example, is a real change, and I'd guess the Android OS would be similar. But if you can adapt to either of those, then the MS OS 'change' is cake.

[dilettante]

Actually most Android programming is done in Java using a portable-optimized JVM and custom set of libraries and services. I'd suspect it'd be an easy change for a C# programmer. C++ can also be used for low-level programming where needed.

[Nightwalker83]

From what I have heard is that development is moving to mobile development! That is unless you are referring to the development of games and a like that I don't think would look good nor run too well on mobile devices?

[dilettante]

Well that's really "been the word" for several years. A lot of people started to shift to mobile development when WinMo 6 came out, only to be kneecapped by Microsoft with the Phone 7 fiasco that cost Microsoft their market share and made developers throw everything away and start over.

Many are very wary of investing any serious effort into Win8/Metro/RT now for the same reasons.

Mobile games are actually pretty hot, though we're talking about a simpler class of games (Angry Birds, etc.) like those on small handheld dedicated game devices.

[jmcilhinney]

There's no doubt that there's a move to to mobile development but I think that the magnitude of the shift has been overstated in some quarters. There are still a lot of applications that just aren't practical on a phone. The number that are practical on a tablet is far greater but even then there are some things that are simply better on a proper PC. I've heard stories of corporates replacing PCs with tablets and then everyone complaining that they just can't be as productive when they're at their desk.

As for the shift from Windows Mobile to Windows Phone, I think that it's rather naive to say that it cost Microsoft their market share. Windows Mobile was steadily losing market share as it was because your average smart phone user wanted, or had been convinced that they wanted, a finger-driven iPhone-style device. If Microsoft wanted to stay in the race then they needed to provide that, and that's what Windows Phone does. Windows Mobile could still be a good platform for business apps and it's a shame that Microsoft didn't maintain it for that purpose. I can only assume that they decided that it would be targeting too small a market to make keeping a second mobile platform alive as well.

Given that Windows 8 now sports the Metro interface, I think that it's a fairly safe bet that Microsoft will stick with that and push it hard for a while to come. From what I've heard, a lot of developers were letting the dust settle with Windows Phone 7 and were waiting for Windows Phone 8 to start development for the Microsoft mobile platform. From what I understand, the Silverlight/XNA apps that were built for Phone 7 will still run on Phone 8 but with HTML5 now being the preferred tool, getting into Microsoft mobile development may now be easier for those who don't already work with Microsoft development tools.