WolframAlpha computational knowledge AI
maximize 2a-(1/cos(a)-1)-tan(a) where 0<a<pi