Dick Eagleson wrote:
1) No spaceplanes.
2) No mid-air refueling.
3) No part-jet, part-rocket crossbreeds.
I find all of the above to be too complicated and with too many possible unrecoverable failure modes.
I also think: no SSTO's.
There's a lot of debate on that, most of which I tried to capture here. There's no settling those arguments without actually building some working RLVs, and even then we'll be arguing over where the dividing lines are.
Like John C., above, I like the 'Rocket Company'-type "right-angle" ascent profile: go straight up on the 1st stage engines, do staging above the effective atmosphere with a vacuum-optimized 2nd stage, then blast sideways on the 2nd stage engines to make orbit.
I've never been a fan of that approach. Your 2nd stage is practically an SSTO. The first stage performance constraint is so rigid that you can't increase it at all to compensate for any problems you have in developing the upper stage. So you've got almost all the difficulty of an SSTO with the extra cost of the 1st stage.
For touchdown, I think either a parachute (round) and Kistler-style airbags or a main engine restart and high-G braking burn near the ground using residual fuel, followed by soft landing on a giant "air mattress" like the "Cloud Nine" units used by Hollywood stuntpeople. Probably have to make it out of Nomex or some such.
This sounds like the original Kistler architecure, which used a trampoline to avoid the weight of landing gear. Having a requirement to land within a few meters of the target to not destroy the vehicle is not a good way to improve reliability.
I like round boosters for aerodynamic simplicity when compared to "flounder-shaped" boosters like the X-33/VentureStar or aircraft-shaped boosters like some flyback designs that have been proposed. I think this would require a cruciform or hexaform or octaform aerospike engine complex with four, six or eight linear aerospikes radiating from a common center on a reasonably large-diameter booster stage to work.
X-33 had its shape because the thrust loads from the engine have to be distributed to the structure as a whole. Putting a linear engine under a round vehicle requires a lot of supporting structure. Multiple lines won't work because you'd lose atmospheric pressure at the intersection point (the rocket plume pulls air away faster than it's replaced, there's no pressure confining your exhaust, and you've got an engine with no nozzle).
the linear aerospike engine built for the X-33 worked fine. The project failed because the rest of the vehicle design goals couldn't be met by Lockheed. I think that's an existence proof for the practicality of the linear aerospike concept.
X-33 didn't disprove the LAS, but it never flew it. They didn't even fire the mini-engine on the SR-71. So it's a good idea but I don't think X-33 added any weight to it.