There's a distinct relationship between how much I've got on at work and how much I feel like coding in the evenings and weekends, and that's not made much better by the nagging feeling that I could be spending my time writing useful utilities for work that I've been putting off instead of messing around with XNA, WPF or something else that's cool, interesting and worth blogging about. I have actually been doing a bit of XNA when the mood takes me, and that includes finally taking a look at the Racing Game starter kit. It's a fun enough game,...

The weakest part of XNA is the documentation, which is understandable since it's a beta. I've pulled together some feedback based on what I've seen and been told so far. All code examples should at least compile The missing class/member documentation is annoying but excusable for a beta. The 2D tutorial is OK, but the 3D one requires that you have too much prior knowledge about 3D. This is a major barrier to actually using the 3D if you haven't come from another system such as DX. The tutorials are OK with the what to...

Last night my brother gave me his comments on XNA after having tried to use it for a full day. The API and usability are very good, but the experience is bitterly let down by the documentation at this stage. The examples contain errors that, OK, I can fix but only because I know enough about how DirectX works in the first place in order to make the fixes but if you don't then you're pretty much stuffed. The rest of the documentation is unfinished and missing pages and methods which makes exploring the API hard if you don't have...

My brother is an animator who dabbles in writing small game type things in his spare time using tools such as QBasic. Before when I've asked him why he doesn't move up to .NET or C++ he's always said that they're too complicated and need too much learning when all he wants to do is play and make versions of Dizzy and Raycasters (the type of engine behind Doom 1+2 for people to young to remember before real 3D). Yesterday I gave him XNA express, my wired XBox 360 controller and showed him the help file. After a bit of coaching about C#...

Part 1 here. So now I have XNA in my hands and I've had a chance to play. Is there anything revolutionary here? No, I don't think there is. Is it a good product? Yes, it's a great product for making games and when the 360 is supported then it'll be something really special. As far as I can see you wouldn't want (or be able to) use it inside of normal applications to provide 3D or the like but if you want to make a game then it's a great starting point. I'll have to code some more in order...

This week sees the release of XNA and while it's not yet ready for the XBox the games you write will work on the PC and so I'm beginning to wonder who is going to start churning out code. The thing is that most people are talking about new developers and new games, but I don't think that we should discount the existing code out thre on the web. The biggest and quickest source of games will be open source projects like ScummVM. It would be great if somebody ported that to the XBox and so I suspect that somebody will do...

With the upcoming release of XNA I'm beginning to wonder where I stand with DirectX. I code with C# and we currently use Managed DirectX 1 in order to provide some 3D and sound related tasks. Lucky for us we didn't go down the MDX 2 route because we would have been shafted, what with it's being canned before the end of the beta and all. What really worries me is that XNA is marketted as a game framework. We don't write games, we write high performance visualisation software and so when we're playing audio we're decoding it from an obscure...

This is interesting. Apparently they have managed to remove all those pesky Sines, Cosines and Tangents from trigonometry. I wonder if the new way will be faster when implemented for things like 3D.

I spent the day today trying to speed up my texture loading for my app as much as possible while, at the same time, trying to reduce the texture sizes as much as possible because the original data (which was a subset of the full data anyway) came on 11 CDs and you can only add so many extra DVDs to a product before it starts to annoy people. I don't have any conclusions yet, but when I do they're going to be so specific to my unusual use case that they're going to be irrelevant for everybody else so...

A couple of times a year we enter a panic because we need to get some software ready for some show or another. This generally means that I end up working of show specific code instead of doing things such as, say, fixing massive memory leaks in my Direct3D texture management. Now that I've got a chance to look at it again I think I may have fixed the problem, but I'm not sure how as all I did was fiddle slightly. I may have had a stupid bug where I was stopping the memory from getting GCd by holding...