I actually wrote some code this weekend in between writing assignments for my degree (Write 300 words on this painting, concentrating on the way the artist has used the elements of the painting to give it meaning: It’s a landscape. And the sky is blue. And I still need 292 words for this section, and that was only one quarter of the assignment. At least the next module is actually on history.) and getting my network working with the fact that I now have XP pro on my desktop machine (IIS and remote desktop. I’m terminal servicing into my desktop from a laptop in front of the TV over a wireless LAN. Isn’t technology great?
So what did I write? I actually wrote some .NET code for a change. My day job is all embedded XP code with serious performance requirements at the moment so I can’t use .NET without serious risk to my severely short project timescale (doable short though, which makes a change in this industry). Does .NET have overlapped file IO? And if not, why not? I can’t write my app without it. Well I took some time out to look at the possible migration strategy for converting one of our current MFC applications to .NET and I have to say that Microsoft are really missing a trick here. I’ve heard rumour that Whitbey will add more support for this, but since that is now still a year away it doesn’t help people like us adopt .NET with our existing applications today. The best I can come up with is to compile with the /CLR switch and have a background project to replace the main EXE of our app with a C# version for possibly the release after the next major release. What Microsoft really need to do is to post a few articles (or make them easier to find if they have already) that explain how to do things such as add .NET GUI code to an MFC view. It’s great that I can flip a switch in the compiler and get access to the whole framework and the ability to call C# classes but the real gains of .NET for our project is the GUI and we can’t just rely on being able to show dialogs. It’s how we’re going to start out I think, but it’s just not good enough for us to make the major leap without major rewriting of our code. If anybody has any idea on how to do what we want then please add a comment to this post.
The other thing I was writing was an app that uses my MSI classes to help take a windows installer file and make an embedded XP component from it. I get the feeling that I’m going to need that tool a lot over the coming months with the number of custom eXP images I’m going to have to make for obscure hardware.
You know what? I’d forgotten how much I like coding C# over C++. I think I need to change some things so that I do it more often.
