I've been seeing a lot of articles about the role of a software architect lately. Not surprising considering the ongoing financial crisis. There's a lot of job-changing going on, and people have to defend their worth to their employers. Naturally, me officially being one, I have an opinion on the subject.

I'm working on some middleware Java EE software and ran into an interesting problem last week. I had a background job that retrieved a list of records from a datasource and processed them. If an error occurred, a status field in the records was set to 'ERROR'. Unfortunately, the Java EE timer service runs in a transaction, creating a very nasty problem.

I've been using Java Swing for over decade now - ever since it was first released in fact. Working with Swing has always been somewhat of an acquired taste. You can do very cool things with it, but you can more easily do very uncool things. When I was neck deep in SwingUtilities.invokeLater synchronization bugs, I'd wish for a better UI library for Java.
Over the years I tried alternatives, such as QT-Jambi and SWT, but they never really convinced me. Recently, I decided giving JavaFX 2 a whirl. I wasn't convinced about JavaFX, especially after Kirill Grouchnikov called it a trainwreck and left the Java UI scene behind. On the other hand, Amy Fowler was on board, and she was one of the core Swing team back in the day.

This article won't be a comparison of the two -maybe I'll do that at some later date- but a code example.

I thought I'd do some mythbusting, for a change.
I found this article on Slashdot the other day. The premise is that 'Splashscreens are sloth'. That people should create better software that comes up instantly.
An interesting rant, but does it measure up?

Each day, In the train to work, I pull out my laptop and work on my own projects. I turn it on, plug in my headphones, and then follows my morning ritual:
Alt+F2, open konsole, open alsamixer, check auto-mute, enable auto-mute if disabled, close alsamixer, close konsole.

I always have regrets about software I build. No exceptions. I've been designing and building applications for over a decade, but I still can't make all the right choices.
I've come to the conclusion that you always make the wrong choice for for each piece of software at some point. This has taught me two important lessons.