So many places to criticize the article, but the one that hit me the most was "Release Early, Release Often" being held up as counter to usability. My experience is the exact opposite; it's not until you've released that you actually know how people will use your product and often too much UX design up-front is just premature optimization.
That's true, but then you have to follow up with users and take careful notes.
The impression I get with many open source projects is that they just push out updates in a vacuum and aren't so interested in "regular" users feedback beyond bug reports and patches.
Some aren't interested, I'm sure. But I think this is one area were better tools could help a lot.
Getting good usage data is hard. Running a good user test is harder. It's my experience that if you connect developers up with good data, both qualitative and quantitative, many of them will start hacking the user experience with the same sort of relentless optimization that they bring to code quality and system performance.
I think there has to be a good solution in there somewhere. Maybe several.
