Programming brick walls

This past weekend I spent a lot of time with my developer hat on. I don’t have much to show for it though. That’s the way it goes sometimes.

My goal initially was to clean up the Firefox plugin and post it somewhere on one of the servers so I could start getting some feedback from everyone. I’m sure there are bugs and I know it doesn’t mean that much that it can run on my Tablets.

At first I figured I could tie the Firefox addin into Community Server (CS), but then I realized how much I didn’t understand about CS, so I decided to go with something simpler. But what? For Plan B I decided I should get the plug-in to work as a SearchTip. Of course, that meant I needed to migrate the code in the current SearchTip (which is written in C#) to the Firefox plugin code (which is in C++). Normally a port isn’t that big a deal, but this was the weekend 🙂

And somewhere around 10 Saturday night I started realizing that my simple Firefox ink plugin was getting pretty complicated. Did I want to do that? Or should I step back and create a simpler ink example that doesn’t involve recognition. (Actually it’s not the recognition that’s the issue, it’s the editing support that’s so crucial.)

So about 11PM I decided I should try a drawing-focused example. I have a little web-based drawing app I wrote awhile back; I could pull that out and get it to support the Firefox plugin. I played with that a bit. Then I realized–yeah, I could do another little drawing sample, but what would make this more interesting is some type of Ajax communication back to the server which could save strokes or allow people to edit each others drawings in real-time. This was a bad idea. Not because it was a bad idea per se, but rather because it was a big idea for 11 o’clock at Saturday night.

I’m not sure how long I spent going down this path, but I do remember that I was back at it Sunday morning before I was awake. Most of the time I dabbled with the Atlas Toolkit, trying to figure out how far I was going to take this approach. I was rapidly learning what I didn’t know, which was a good thing except that my “didn’t know” column was growing rapidly.

About breakfast time I realized there was no way I was going to get this done in a day–nor two days–nor three. So I demoted the project from a concerted weekend effort to a coding-while-watching-TV-in-the-evening project.

Since I was in learning mode at this point, I decided to work on some WPF stuff for the rest of the day. There went the rest of the day. First, I listened to some great PDC05 presentations on various WPF projects here. First, let me say that these rebroadcasts of the PDC05 presentations are phenomenal. I hope Microsoft does the same thing for TechEd. I learned a lot. There’s nothing like being right at the conference and being able to ask your questions in person, but it’s great to be able to pause and rewind what’s being said too.

All wasn’t perfect, however. WPF is still being designed. And all of the samples shown in the PDC presentations require at least a minimal set of changes. The best list of what changes you’ll need to make to “older” XAML samples is here.

Unfortunately, this list didn’t help me get the one PDC sample working which I was most interested in: This circular list box which illustrates how to specify custom layouts.

Actually, there’s a lot of XAML stuff on the web now that’s broken. This is part of the fun in working with software that’s not fully baked. It is very interesting to see the progression though. You can see how the XAML “model” is becoming more complete and complex. I’d go so far as to predict that by 2.0 XAML will be looking more like a full programming language than an RC file. We’ll have to see. I guess it depends on how stubborn the team is. In something as crucial as a language, design stubbornness is a good thing. Sometimes engineering stubbornness, however, can’t overcome market forces. I’m not placing any bets yet, but I have my eye on market forces winning this one. 🙂

In the meantime, it sure would be nice if the XAML team could update its posted samples. Part of the challenge will be that older blog posts are no longer accurate. This is kind of like the old MSDN problem. If there was an infinite amount of time it would be nice to update all the old posts. Next best would be to link to an updated version of the code (Maybe the community can help out here.) Short of this, a note would be nice indicating that the code requires modifications (simple or not).

Anyway, I’m glad it’s Monday. I can get back to my real work. 🙂