The distinction I had in mind was between operations in the Earth-Moon system versus operations beyond our front yard. Many people have pointed out that the delta-V requirements for getting to Mars aren't much greater than for getting to the Moon, but it's the travel times that make the dramatic difference (for the near- to mid-term, until we get some really powerful engines going). Travel time primarily impacts life support requirements, with big implications for mass and lesser implications for things such as communication gear.

Travel time does also affect the choice of propulsion system, I think. High-energy/short-duration chemical propulsion systems are great for getting around the Earth-Moon system; while low-thrust/long-duration technologies like ion engines start to pay off on long duration missions where small continuous acceleration has huge cumulative effects. ESA is going to launch a small probe to the Moon using ion engines, and it'll take months to get there. That's a fine way to save money on an unmanned spacecraft, but using that technique for a crewed mission would be insane--the greater mass requirements for life support would entirely negate the benefits, if not make it completely impossible.

In all that, I'm thinking of the relatively short-term, creating a system based on technology that we either already have or is under development now. No question that in the long run we need the energy-density of nuclear propulsion. I may sound like a liberal over on Current Events, but I'm quite comfortable with the use of nuclear power in space. We can't get there from here without it.

Here's an earlier concept for a modular deepspace craft built around a keel:

Who wouldn't recognize Discovery, eh? Sir Arthur really knew his stuff.