Closing in on 11am. The day so far: Fixing flakey machinery, fixing flakey documentation, trying to adjust weird requirements. Sometimes if the solution isn't likely to change, is it a valid approach to try and change the problem? (Meanwhile, a friendly wind's brushing through the trees on the other side of the road, and a dim moon seems stuck in hazy clouds. A wave of warmth flooding through the city, but its taste is already late summer in every nuisance.)