I like the idea of iterative development: you integrate a new thing, you take some feedback, you tweak the new thing, you tweak the whole site and you repeat.

After adding discussion forums and blogs to FanConcert, the site has taken on a new personality. This is exactly why I didn't want to tweak too much too early: I add a simple feature like forums and the whole mood of the site changes.

The trick to software agility, I think, is to not get married to your small tweaks. It can be hard to do but the greater good of the site is more important -- and that's integrating new, big features into the site.

Every time a new feature is added it's good to step back and see how that effects the site as a whole. That's what I'm trying to do with FanConcert: see the forest for the trees.

Given that I'm under self-imposed time constraints, I'm using my time to integrate more of these big features into FanConcert than constantly tweaking the current state of the site.

After all, a new feature could come along and wipe out all of those tweaks because it needs to. Overall that can translate into wasted time spent tweaking too much. It's better not to be emotionally attached to those tweaks, it makes them easier and faster to completely nuke. Buh-bye!

It's hard though. I want to make the site look good (and navigate well) right now but I have to balance that with some pragmatism. If all I do is tweak this site is never going to work for people ... and I think that needs to be the bottom line.