Also added a preference to keep the window above the screen saver
Anyone who truly understands UI design realizes that every preference option is an admission of defeat: it's there because you couldn't just get it right the first time.
Sure, most UI behaviours are unsatisfactory because they're too far from the brain. Sloppy focus is a poor man's focus follows brain, and this sort of pref should instead work by measuring your expectation when the screensaver runs and either disappearing the clock or moving it to the front as appropriate. Likewise, when I hit Ctrl-T in the wrong window, it should affect the right window anyway, and when I type !$ it should mean the last parameter of the command line I was thinking about, not literally the most recent and probably irrelevant command. Babbage was wrong to criticise users who asked if the computer would give right answers when asked the wrong questions, of course that's what users want, who wouldn't?
For a really cute example, get towards the end of Portal 2. The game suddenly puts the player into danger, giving them only a second to use a portal to escape certain death. In play testing about 50% of people panic and set the wrong end of the portal, but they aren't sure and so the only feedback they have is that they died. This is frustrating and doesn't add to the game. So the game has a special case, for that sequence - any portal you create will always be the right one, no matter which colour you actually asked for. Everybody gets to feel like they got the right answer even if they didn't.
no, he's quoting jwz.
Yes, I took that as read. I was just explaining how it becomes necessary to admit defeat in these cases and thus why Jamie's earlier words, while perhaps a little ironic in such proximity, are not hypocritical.