XNA and DirectX (outside of 3D) being phased-out

[PTNL]

I caught this blog post recently from Promit Roy, a Microsoft MVP on XNA and DirectX. In it he goes through the official announcement from Microsoft, and comments on some observations.

A fair number of developers in this forum primarily work a couple layers above C/C++/bit-fiddling. So I'd like to hear from those that are more involved with the DirectX, OpenGL, XNA, game development or unmanaged codebases, and could be affected by the announcement -- at least to some degree.

Here are some possible questions to help with responses:
- What is your development background? How many years, and in each area?
- Can you summarize, or at least generalize, what you have been working on -- either on the job, or during off hours?
- What investment have you put into XNA and/or DirectX? Did you publish anything?
- What level of impact does the announcement have on you?
- If you had also seen the writing on the walls of this announcement, what tipped you off and when? What technology movements did you consider, and (if a decision was made) what was decided?
- What other relevant technologies are you currently focused on? What made you choose them?


Edit: My hope is that this thread becomes a a chance for professional devs in this forum to chime in on a topic, if it directly affects them.

 

[wonderfield]

He makes some interesting points. The DirectX ecosystem has been fragmented with new APIs intended to replace older DirectX APIs, and the widely-held opinion among developers is that these new APIs aren't particularly good. XAudio2 isn't really a great improvement over DirectSound, and the raw input APIs aren't nearly as pleasant to work with as DirectInput (which is one of the more well-designed systems Microsoft's produced). XInput only supports the Xbox controller, and the story on building desktop games with controller support without the old DirectX SDKs is fairly grim. It's still doable, but you need to reach a little further into the toolbox to get it done, which diminishes the appeal of Windows development somewhat.

It's not particularly clear why some of these changes are taking place, but it does seem fairly clear that Microsoft's interest in Windows as a gaming platform has dwindled from where it once was. The strong momentum that once was is now almost entirely absent, and we're seeing only handfuls of improvements to Direct3D as time goes on. The rest has stagnated entirely.

That said, development of desktop games is still supported. The integration of the SDK with Visual Studio makes things pretty pleasant for those working on open-source projects, actually: if you stick with what's available, you can just release project files and the source and other developers can just load that into VC11 without any further configuration, and it should all just compile and work. It's not a huge win over the way it was, but it's still a win.

As far as XNA is concerned, the new avenue for that type of development is in WinRT. That's going to bother many who want to target older Windows platforms, but those platforms are still targetable through C/C++/DirectX, and it's probably the path many serious developers should be leaning toward anyway. Things get harder there, but things get better, too.