\(\displaystyle Another, \ easier \ way, \ I \ think, \ to \ wit:\)
\(\displaystyle Prove \ that \ 2tan^{-1}(1/5)+sec^{-1}(5\sqrt2/7)+2tan^{-1}(1/8) \ = \ \pi/4 \ rings \ true.\)
\(\displaystyle This \ can \ be \ rewritten \ as: \ 2tan^{-1}(1/5)+2tan^{-1}(1/8)+tan^{-1}(1/7) \ = \ \pi/4.\)
\(\displaystyle Now, \ let \ 2tan^{-1}(1/5) \ = \ \alpha, \ \implies \ tan^{-1}(1/5) \ = \ \alpha/2, \ \implies \ tan(\alpha/2)=1/5, \ \implies \ tan(\alpha)=5/12.\)
\(\displaystyle Also, \ let \ 2tan^{-1}(1/8) \ = \ \beta, \ \implies \ tan^{-1}(1/8)=\beta/2, \ \implies \ tan(\beta/2)=1/8, \ \implies \ tan(\beta)=16/63.\)
\(\displaystyle and \ finally, \ let \ tan^{-1}(1/7) \ = \ \gamma, \ \implies \ tan(\gamma) \ = \ 1/7.\)
\(\displaystyle Hence, \ \alpha+\beta+\gamma \ = \ \pi/4, \ implies \ \alpha+\beta \ = \ \pi/4-\gamma.\)
\(\displaystyle Taking \ the \ tangent \ of \ both \ sides, \ we \ get: \ tan(\alpha+\beta) \ = \ tan(\pi/4-\gamma).\)
\(\displaystyle With \ a \ little \ algebra \ and \ appropriate \ identities, \ this \ reduces \ to \ 3/4 \ = \ 3/4, \ QED.\)
\(\displaystyle Hence, \ the \ equation \ (above) \ is \ true.\)
\(\displaystyle or \ tan[(\alpha+\beta)+\gamma] \ = \ tan(\pi/4), \ \implies \ 1 \ = \ 1.\)
\(\displaystyle or \ for \ the \ purists, \ let \ \pi/4 \ = \ \theta.\)
\(\displaystyle then \ \alpha+\beta+\gamma \ = \ \theta.\)
\(\displaystyle Ergo, \ tan[(\alpha+\beta)+\gamma] \ = \ tan(\theta).\)
\(\displaystyle tan(\theta) \ = \ 1, \ \implies \ \theta \ = \ \pi/4, \ which \ is \ what \ we \ wanted, \ QED.\)
\(\displaystyle Note: \ To \ make \ this \ problem \ more \ interesting, \ it \ should \ have \ been \ written \ the \ following:\)
\(\displaystyle 2tan^{-1}(1/5)+sec^{-1}(5\sqrt2/7)+2tan^{-1}(1/8) \ = \ \theta, \ solve \ for \ \theta, \ (without \ a \ calculator) \ 0 \ \le \ \theta \ \le \ \pi/2.\)