How can we get to the Moon for cheaper?

1. You need ISRU production of propellant. And no, I don't mean processing regolith, I recommend processing the lunar ice, or taking a look at those mysterious active volcanic vents that apparently put out H2O, CO2 and other useful gases. Do this with robots, make a fuel dump on the Moon. This is a technology demonstration and development program.

2. Once you have working ISRU you don't need to send separate lunar ascent and descent stages like Apollo, just refuel once you're down. It's only 2km/s of delta-v.

3. Stop talking about zero-g propellant transfer.. acceleration settled propellant transfer is TRL 6 and the slight propellant penalty is cheap.

4. Put a propellant depot in LLO, now you don't need to carry propellant for your descent stage up from earth either. You refuel in LLO before making descent and refuel on the lunar surface before making ascent.

5. Tug propellant from LLO back to L1/L2 and LEO. The goal should be to reduce propellant launches from Earth to zero.

6. Big roomy capsules are a luxury that should come later.. they should stay in space and be reusable, either by getting around completely propulsively or by using aerobraking.

7. If you go with small capsules, do you need a separate lander? Initially there's mass savings to be had but specialization of landers that remain in LLO or on the lunar surface will eventually win.

To summarize, ISRU and on-orbit propellant depots make it cheap, do that now. Reusable in-space vehicles specialized to a particular leg of the trip are the ultimate way to lower costs.

Some more thoughts

Some things immediately spring to mind when I hear people talking about lunar resources:

• The only resources that have been characterized are regolith and rocks of the landing areas of Apollo. If you want to use lunar ice you need to characterize it first. So are you going to do that with humans too? If not, sounds like NASA should be working on a robotic precursor mission right now.
• Maybe you could make some educated guesses about what you're going to find.. solid ice, dirty slush, wet regolith, dry regolith (but more wet than the stuff Apollo brought back), etc. So you could immediately start working on a demonstration processing system, and focus that effort once the characterization information comes in. Either way, eventually that demonstration is going to need to fly to the pole and show that it can do the job.
• Assuming that demonstration works just fine, next you have to scale it up. Around this time people are going to be pushing for that scaled up project to be human tended. I'm actually of the opinion that you should avoid that at all costs, but good luck with that.
• So how do you ramp up the utilization of lunar resources? The traditional viewpoint is Lunar Surface Rendezvous.. aka, refill your single vehicle on the lunar surface and then fly back to Earth. As interesting as this is, I don't think it works. Better idea: a reusable automated lunar lander. Now you only need a lunar orbit capable vehicle to get your humans there and they can take the lander down, and back up, and none of that fuel needs to come from Earth.
• What's the maximum leverage you can get out this architecture? Well, you can at least get Earth return propellant into LLO or even the Lagrange points, so you don't have to take that from Earth. Some would say you can send propellant back to LEO and now you have a complete cis-lunar space infrastructure. I'm not convinced of that though, I think you at least need to launch propellant for LEO->L1 from Earth. And if you're going to mass fuel up at L1 then you might as well build a permanent space station there.

Now figure out how to fund all this.. and stage this stuff so each incoming Administration doesn't feel they've had the "vision" dictated to them.

But I don't think you are asking the right question. The right question is always:

Why?

What is the purpose of going back? If the answer is "exploration" then you've just begged the question.. What is the purpose of exploration? The purpose has always been colonization, and if you haven't got that in mind, you'll go exploring the wrong things.

For example, Zubrin will tell you Mars is a much better target for science because there could be life still on it somewhere and finding that life will be much more interesting than publishing some papers on lunar geology. I don't disagree, but what's that got to do with colonization? If anything, it adds weight to the "let's not contaminate Mars" arguments that are against colonization.

So what should we explore on the Moon to support future colonization? And no, I don't mean "explore" in the technology maturation sense of the word, although that's important too. I mean, what should humans be sent to look at? Simply:

* Resources. Where are the volatiles we need? Where's the ice, the vents of useful gases, the metal deposits? There's only so much exploration that can be done with robots, and that exploration will be all used up in step 1. The resources they find will not last forever, they might not even be sufficient for more than a dozen expeditions.

* Sites. Where do we build? The simple fact is, the Moon is a harsh mistress. Bubble domes are sexy but I don't know how they provide radiation protection. Any sensible Moon colony will be built underground, so that's where we have to go exploring. That means caves, lava tubes, and other natural formations. The goal should be to identify how much heavy equipment will be needed to build a base.

There will be plenty of opportunities to do science along the way.