xna dead?
I wouldn't learn DX or XNA. I'd find a nice "engine" and use that instead. It'll save you a ton of time having to implement many game engine features that DX or XNA don't do for you.
I see so many of these XNA Dead threads, and really, I feel that even if there is no support for XNA anymore, there is still such a huge advantage to making a project with XNA if only for the fact, that you are learning so much in the way of game programming, that the experience is well worth it.
I started using XNA a few years ago, and that led me to learn C# (since I was only C++ before that) and since learning C#, I now have a Java class, and I feel really comfortable with the managed sort of langauges. Also, my XNA classes and functions I have ported to C++/OpenGL, so though I dont use XNA anymore, there is nothing that I have lost in using it, but only gained.
So in conclusion, it really doesnt matter if XNA is dead or not, just program, and learn, and have fun.
Time spent with XNA now and even for some years into the future isn't wasted. You'll learn a good deal about programming in general, how to organize larger and larger projects, and all the math you need to make a game. All of that stuff transfers 95% to any other 'curly brace' language -- C++, C, Obj-C, Java, etc.
If you're really concerned about long-term platform support, then I'd be more wary of XNA. If you look at what Microsoft is pushing now, its all native code, not C# or XNA. WinPhone8 supports native apps, WinRT (Surface on ARM) supports native apps, Win8 (of course) supports native apps. They'll run managed (and even XNA) apps too, but Microsoft's focus right now is on native code.
C++ is still a difficult language to cut your teeth on, but C++11 makes it a lot nicer than it used to be. That said, compiler support from Microsoft is only partial now, and it'll probably be another year before the support level is good, and furthermore, it'll probably be another year before most of the "must-have" C++ textbooks have been updated to reflect C++11.
I'd say that your best bet now, if you're already comfortable and productive with C# and XNA (use 4.0), is to continue with those for another 9-12 months, and wait for the next version of Visual C++, and for the books to catch up to the standard. I believe you'll find that to be the best use of your learning time. You shouldn't worry about trying to learn the latest, greatest thing so much, because those things (whether they're APIs, languages, or general practices) tend to come and go. The things that last will come up no matter what tools you're using -- sure, the exact details of those things change over time, but the general "shape" of them stays the same.
XNA is not dead, I still currently use it and it'll be in development for more years, they still use it for indie games on 360, and most arcade games. It won't die any time soon.
Just to clarify, development on XNA itself has stopped, but development with XNA is still active. It still has a good community built up around it and many people are actively developing games using XNA. This community will continue to exist for at least a few more years.XNA is not dead, I still currently use it and it'll be in development for more years [...]
I had something similar happen to me. I spent more than half a year learning Directx9 with the fixed function pipeline, blissfully unaware of how deprecated it actually was, and I wasted my time trying to memorize the graphic syntax. However, it wasn't entirely wasted. I learned a lot about how a graphics pipeline works and how games are made :3 I had a much easier time learning directx11 then I did the first time.
The only thing not worth doing is asking if it's worth using still.
It's not dead yet, doesn't seem like it will be in the near future and would be plenty useful for a long time even if it did stop getting updates.
If you want to learn it, go learn it.
