DirectX and XNA Follow-up

I wanted to clarify and respond to a few things regarding my previous post about DirectX and XNA. First a quick note: the very long standing DIRECTXDEV mailing list is being shut down. Microsoft is encouraging a move to their forums, but in case you’re fond of the mailing list format there’s a Google Group that everyone is shifting over to. If you’re reading this blog post, you’re probably interested in the subject matter and so I highly encourage everyone to join. MS Connect’s entry for DirectX is being discontinued as well, so I’m not sure how you report bugs now.

Something on the personal front: I got a few comments, directly and indirectly, about being an MS or DirectX “hater”. Good lord no! I adore DirectX, XNA, Windows, and Microsoft. I criticize because I want to see these technologies thrive and succeed. I’ve been doing a lot of iOS/Mac/OpenGL work lately and from a technical standpoint it’s absolutely miserable. I miss the wonderful Microsoft world of development. But a lot of what I’m hearing in public and private worries me. My alarmist approach is designed to bring attention to these things, because oftentimes the development teams live in a bubble separated from their users. (Hell, I can barely get in touch with people using SlimDX.) The XNA team, for example, are terrible at communication. It’s exasperating, because the direction, plans, and schedules are completely opaque — even to those of us who have signed an MS NDA and are ostensibly supposed to see this information early. (We don’t see jack shit, by the way. Exasperating.)

Second, no I do not think DirectX is dead. No I do not think everybody should switch to OpenGL. From a platform and technical standpoint, we’re probably as much or even more of a commitment than in the past. What’s bothering me is the pathetic way that community, documentation, etc is being handled. Look: Where is the DirectX SDK? I don’t know when that was published, but it appears that nobody noticed it until very recently. How would we? I don’t wander MSDN online at random. And it’s been placed right next to this comically worthless page. This is the kind of developer support I’m complaining about. Nobody outside MS understands what is going on in Bellevue, and I’m getting this worrisome feeling that nobody inside understands either. The DirectX SDK wasn’t just a diverged way to deliver support for a core platform technology, which is what seems to be driving the current decision making process. It represented half a gigabyte of commitment to the developers who arguably make the entire Windows ecosystem compelling to a consumer. “Developers developers developers!” “Yeah, what’s up?” “Uhhh, hi?” That’s what it feels like. MS wants developers around, but they forgot to put any thought into why. (Here’s a hint guys, ask DevDiv. They seem to still have a clue, C++ team aside.)

And in the other corner, we’ve got XNA. Whoo boy. There’s no point to sugarcoating this, although I’ll probably ruffle some feathers: XNA 4.0 is garbage. It exists to support two things: XBLIG and WP7. XBLIG is a joke, so really the only productive arena remaining is WP7. It’s sad because XNA 3.x was actually an excellent way to do managed development on PC. 4.0 introduces a profiles mechanism that is focused specifically around the Xbox and WP7, and produces a stunningly foolish situation on PC where DX10+ hardware is required but none of the new features are supported. Worse still, that was a few years ago. Now we’ve got DirectX 11 with Metro coming down the pipeline and XNA staring blankly back like the whole thing is a complete surprise. It tends to raise some questions like, is XNA 5 coming? Will there be DX11 or Metro or Windows 8 support? Is anybody even listening? And the answer we got was, and I quote: “We’re definitely sensitive to this uncertainty, but unfortunately have nothing we can announce at this time.”

My judgement on that message is that XNA does not exist on PC. I’ll say the same thing I said in 2006 when XNA was first revealed. Treat it as an API that happens to run on Windows as a development convenience, not as something it’s actually meant to do. The computing world has moved on, and if Microsoft can’t be bothered to bring XNA along then that’s just something we have to work with. If and when XNA 5 is announced, then we can go back and take a look at the new landscape.