Metals have important advantages as mirror substrates. Existing metal mirrors exploit the bulk properties of the material for mechanical and thermal purposes, and some mounting details and useful optical features which are very difficult with glass can be more readily produced in metal. In the near future, metal substrates promise to extend the possibilities for large active optics (with control of the optical figure), on account for their mechanical properties and methods of attachment. Looking further ahead, enhanced active optics will make it possible to produce large non-axisymmetric or variable optical figures. I suggest that such optics will open new dimensions for the telescopes of the future. One potential example is a two-aperture telescope with coherent imaging. Its elements are off-axis Cassegrain telescopes. The two beams combine (if required) in an optically ideal manner and with no obstructions or further reflections. Other suggestions for further study include the reflecting Schmidt camera, beam-steering systems and systems with a non- circular field of view.