red echo

Work continues to soak up most of my time and energy. We are pulling into the final crunch now, and it feels good to see the details finally coming together. We've been working on REALbasic 6 for about a year now, and while that's actually on the short side of normal for an upgrade of this magnitude, it has come to feel like a long slow haul. I went from work on version 5.5, where the incredibly tedious and frustrating process of adding Mach-O support eventually crowded out all but a handful of the fun and interesting language features I had hoped to include, into the version 6 process which was focused mainly on reimplementing work I did in 5.0. You can imagine that the personal satisfaction level in such a venture would not be particularly high.

There are good reasons it had to work out this way, and it hasn't been a complete loss, since a piece of code always comes out better the second time around. After you finish a big project you always end up with a bunch of ideas for design changes you wish you could have made way back at the beginning. While I'm frustrated that I have to scrap the entire debugger I wrote in RB 5 not quite three years ago, I do have to admit that there were a lot of things about the old code that were painful and ugly. The newly rewritten debugger is definitely a better piece of code, and it will be much more pleasant to maintain over the next couple of years. I do wish there had been time to actually add some new features, though; the new debugger just does the same job the old one did, with a polished-up, more user-friendly interface.

Still, two years after its release we have barely begun taking advantage of the new compiler engine's power, and it frustrates me to see so much of my old work sitting idle. We have added many new language features, but for the most part they have been small, simple additions, often things I have worked in through the cracks in the schedule. I want to do big things, damn it! I have plan after plan in my head, all queued up, waiting until I'm done with all this boring but essential stuff and have time to really dig into the compiler again. I probably won't get to do much of it right away, but I can't even get started on anything fun until version 6 is out the door. I want to get on with it already.

Well, who knows what's going to happen? I was skeptical about the all-new REALbasic 6 interface at first, but now I really like it, and I'm convinced this is going to be the most significant new version of REALbasic ever - even a bigger deal than REALbasic 5, which included my brand new compiler engine. We're adding support for Linux - you could build Linux applications in 5.5, but now you can run REALbasic itself on Linux, and that's a completely new world for us. None of the star features in this version have anything to do with me, but I guess I can live with that; the work I've done in 6.0 isn't particularly visible to the user, but it lays the foundation for some important long-term architecture development, so it has still been worth doing. It's just that I'm a compiler guy, and I want to get back to the compiler work; that's what I'm best at, and that's what I enjoy most. I'm looking forward to getting back into it again.

Tonight was music practice: more Bach, more work on "Keep breathing", some playing around with new fragments I'd composed in the last couple of weeks. Sadie and Caroline really seem to like what I've done, which feels good. I haven't had enough free time to finish any new songs in the last month or so, but the creative energy is still there whenever I have time to let it out, so that's alright.

Yay, winter is back - it's COLD out.

Looks like the “first snow of the season” is going to be the only snow of the season. What I thought was a little break in the cold weather turned out to be more or less the beginning of spring; here we are in the first week of February, and it's so warm I had to take the lining out of my motorcycle jacket. The weather could still turn back around, I guess, but the ski season is pretty much toast as far as the Cascades are concerned. Oh well, maybe I'll talk a bunch of my friends into heading up to Whistler for a weekend.

Work, work, work. It's that time of the year again, and I'm not getting much done but write code. At least I'm making progress, though: I finally solved a problem with the debugger code that has been deviling me for months. This bug grew so tiresome that I found myself taking any excuse I could find to push the debugger further down the schedule and work on something else; but deadlines loom steadily larger, and I finished the other remaining item on my to-do list last week, so there was really nothing for it but to get the job done at last. While setting up yet another round of tests this morning, I triggered an obscure bug in the listbox control; desperate for any excuse to procrastinate further, I dug in and fixed it. Upon returning to the debugger code, it occured to me that I had just found the source of the problem: under the surface, the bug I had spent all this time thinking about was really the same kind of bug I had just fixed. Ten minutes later, I had tracked it down, and the solution was obvious: a minor tweak.

Well, then. I don't think I want to dwell too closely on what that anecdote might say about me, or about the art of programming; but that's what life on a large software project is like sometimes.