Slowly pulling out of a terrible weekend spent in the depths of some kind of stomach-virus, in which my input was redirected back out and something decided to cat /dev/null through my output, culminating in a lingering weakness of the muscles and throbbing in my skull, and I realise that I’ve lost focus/motivation on just what I was doing the week before, or even what I’ve been doing this year.
I know I had built up a bit of momentum on simple-jndi. It needs to be a JNDI independent bootstrap database, rather than trying to solely be a implementation in itself, and if I can find a memory JNDI impl that works as I assume it should, things would be good. Robert’s GenericContext (currently found inside simple-jndi) seems to do most of the job, just needs a bit of bugfixing. The Apache Directory Naming surprised me by not liking how I was trying to use it, so I need to hassle Phil to see if he has a clue why. Chances are I’m being a moron.
Finding motivation to get Commons Lang 2.1 released is a bit of an uphill struggle. There’s some nice stuff in there. Util code is a lot of fun to write when it raises itself above the simplistic level and I had a lot of fun writing the DurationFormat class; an attempt to provide the basic functionality you want when wanting to output a bunch of milliseconds as a duration without having to move to a much larger library (ie the very good JODA). StopWatch, which was originally just a play until a co-worker kept using it, is tighter in its validation now. Calling stop() twice will result in an IllegalStateException etc etc. To be honest that’s all that jumps to attention right now, but bear in mind my skull still feels like something is trying to drill its way out.
At the Jakarta level, the download pages seems to be working okay. I’ve not heard any bile 
Sebastian Bazley is doing some great work at improving it to remove many of the rough edges I left in. I want to improve the javadoc management at Apache, and I guess the main focus I should have is on easing more CVS->SVN migrations on through.
ReportRunner’s gone down very well at work. The only black cloud for it is the existence of xReporter, who finally got their demo working again and looks pretty good. Eric and I sat down and looked at its site a few weeks back before the demo was there and decided that RR still had a unique approach/feature-set over xReporter that we should continue it, though I do suspect that we could implement those on top of xReporter. Given that we were ready to goto production with RR the real question was whether we thought we should start planning a migration to xR.
Reading blogs about Ruby on Rails is interesting. I need to ask dumb questions a bit to really understand, but it doesn’t seem like anything language specific. Any of the scripting languages could do the same thing (and probably are) while Java could do the same either in a typed way or via untyped maps. I assume somebody is doing that too. My understanding so far is that it’s a simpler Struts MVC + a dynamic DB model that can be overriden. Throw in code-generation as mis-direction (all the examples seem to generate 2 lines of code each time), the natural advantages of a scripting language and everybody’s wet.
Still, it creates a nice bar for the Struts/JSF/WebWork/Tapestry et al to work to. I’ve only really used Struts on that list and it’d be nicer to have something with less legwork (ie automatic Forms and standard Actions that may merely be plugged in - I presume these exist too, they always do).
It’s hard to remember what else I was thinking, pre-weekend of pain. Saturday morning was spent improving the backups from my servers to my home server, so at least that got done. Otherwise it’s tempting to spend the next month vegging until people start asking why I’ve not done X or Y.
