We just got back from doing some solar thermal prototyping in the Arizona desert. We are testing our ground based solar concentrator with 2 dimensional curvature rather than 3d; the 'trough' rather than 'dish' geometry. The ground-based trough is still the spherical geometry that we've used in our popcorn making basins, so the reflector itself doesn't need to track the sun to stay focused.
Linear scalability is the main reason the trough is attractive; a trough can be extended in one direction without the hole getting any deeper. The lamination of reflective material is also simplified, as is the tracking of the sun. Results have been underwhelming so far though- it seems that trough is much more persnickety about the accuracy of the surface. Nearly all trough-style installations already out there in the world use a vacuum tube receiver to achieve usable temperatures, so we will be integrating some of these this spring at our Kelder's farm experimental site. Unfortunately the vacuum tube receivers that are available are just that- tubes- which means they are optimized for a parabolic line, rather than the focal plane that our spherical optic generates. Anyhoo, it was great fun to do some experimentation in the desert, I love digging circular holes.