State of the project

My team's product development manager put in his resignation notice and will be leaving the company at the end of the year. This leaves 2 heads-down coders on the project team now, and one part-time (our VP of development, who says he would prefer to code all the time--though I would prefer his salary I'm sure! ). Since our PDM also put in lots of hours writing code, his leaving is going to impact our schedule pretty dramatically.

In some ways I'm concerned about my job security, but not overly so. There are some bright people in the company and if it turns out that this venture proves to be too costly, I'm sure I will be allocated into some other project in the same technology plane. Of course, it doesn't hurt to be prepared so I've begun to adjust my thinking that I will be pulling consistent income forever, and have started to count my pennies a bit. Outside of buying a new video game and steering wheel attachment for my computer, of course, as those are necessities of life. . Oh, and this whole Christmas thing...

The work I'm doing is getting more interesting as I see an end is near to the particular tasks that I've been tackling for the past month or so. I have a good handle on the requirements and have gotten into a bit of a flow working with LLBL Gen, our object-relational mapper. Database queries have taken on a whole new meaning and in some ways there's a bit of art to convincing the thing to work. As is the norm, knowing what you're trying to do is sometimes the challenge--implementing it once you have a plan is not terribly painful.

Visual Studio 2005 is working out fairly well. There is a binding bug but I've adopted a workaround from Rocky Lhokta that seems to be working, although my combobox population seems to be on a one-change delay in some cases. Like the changed event is firing too late, or maybe I've subscribed to the wrong event. The new features of the designer are pretty cool, though I've found the table layout panel to be wholly unusable for my purposes, as the designer and actual view are very much not in sync. I would like to use this panel for its dynamic scaling abilities, but MS will need to fix it first.

Finally got our automated build process with Cruise Control .NET up and running again (I think...). There was a .csproj.user file uploaded to our Subversion repository that had a hard-coded reference path in it, so that was messing up our build (since the build server uses a path on a different drive letter). Since NAnt wasn't working with VS2005 with the initial RTM, I implemented msbuild as part of the CC.NET process. At some point, I'd like to either convert back to a working NAnt version, or implement some plugins or executable pass-throughs with msbuild to do some of our more complex build tasks that we're currently skipping.

We converted over to using web services (from .NET remoting, as we want to use a more standard approach that will be upgradable to the next gen technologies that MS plans for remote access (Indigo and all that jazz)) last week, and all my LLBL fetches and saves broke. I believe I've managed to repair most of them, and in the process have streamlined the prefetches to only get data needed for each page. On one hand, this change limits the amount of data pulled--on the other creates another round trip on page load which didn't exist previously. I think the new method is more efficient, but I haven't run any tests on it.