Always dive down into a problem and get your hands on the deepest issue
behind the problem. All other considerations are to dismissed as
"engineering details"; they can be sorted out after the basic problem
has been solved.
-- Chris Crawford