Seems very very good really. It took a very short time to get a load of transulcent sprites bouncing round the screen, and I could get similar numbers moving about before the frame rate started dropping as I can with C++ and Direct3D as far as I can tell.
So I guess I have to really ask whether it is worth continuing to develop Udo in C++ or should I start it again with XNA?
I've been messing about today figuring out how to load my sprite files onto Texture2D surfaces. I'm unsure whether or not implementing the texture atlasing system will actually affect performance (to be honest, I never actually tested the performance benefits under Direct3D either, I just assumed) so probably need to do some testing to see if it is better than just creating a texture per image.
Otherwise I'm actually quite interested in writing a full size game in C# instead of C++. I'm growing quite attached to it.
Job-wise, few irons in the fire at the moment but no real news.