Be careful not to prescribe technologies that carry cost penalties.

For instance, your statement earlier that we need to develop fully-reusable technology with a 1-week turnaround time is likely to be a VERY expensive route to space. You need to build hardware that will survive the vibration and stress of multiple missions, and be able to strip-down, inspect, and rebuild this in a week? (or have multiple vehicles overlapping).

As we saw with the shuttle, re-entry is a complex business and is expensive in terms of the TPS system and its inspection overhead, which is mission-critical.

In many ways, it is cheaper and more efficient to have a Henry Ford assembly line stamping out much cheaper throw-away bodies. This is the approach the Russians have taken in many ways, which is why they can get a paying customer to space for a mere US$20 million.

If you only have a design life of one mission, everything can be made lighter, thinner, cheaper, and testing made simpler - you are always testing new components, not dirty used ones, and they have to pass less-severe inspection. You completely eliminate the need to maintain an incident log of the adventures of each component on each flight. You build it, kiss it goodbye, and collect statistics during test flights to see if you need to build it stronger next time.

Every saving in component mass multiplies exponentially into a saving in fuel and cost of the whole package.

Concepts such as top-end steering allow you to concentrate the avionics in a compact (and potentially recoverable) package at the top of the stack. or even integrate them into the final payload in a way that gives you more options (big steady gyros and steering motors can double as a long-term steering and stabilisation system)

Gary likes the idea of a biamese fly-back system. I can see some merit in it, providing that both boosters are released at low enough altitude and velocity to avoid the need for extensive TPS on a winged vehicle. IMHO, such a combination is criminally stupid, as Columbia has demonstrated.

IF both boosters can be released and fly-back under "mere" aerodynamic stresses, then they can be built for a long service life. (but I doubt it was cheap to overhaul an SR-71)

I like Gary's method of cycling the airframes from high altitude use to lower altitude use as they age, but I disagree with the concept that ANY of them should actually re-enter.

I think a cheap and dumb third stage should provide the punch to get out of the atmosphere, and be expended in the process. That gives the least weight penalty. (although you might finesse the operation with a more expensive and complex higher-performance third stage that would still yield nett benefit, despite being more costly, by allowing higher payload per launch).

And build the whole system BIG. *

Build it local to the launch site (no more shipping of pieces across the countryside and being limited to an evolved ox-cart loading guage).

Build it somewhere where recovery of intermediate stages is easier. For example, launching from Baja California across the Gulf of California (for immediate post-launch ground safety) then gives numerous opportunities to fly intermediate stages down to a series of landing sites through Mexico and Texas, without any cities being on the direct launch path.
Or, a Central America launch across the Gulf of Mexico, with landing options through the Carribbean.

Singapore would be good - equatorial, with all the Indonesian archipelago as landing sites.

The value of this is that you only need limited cross-range capability for the fly-back stages, so they have smaller wings, less mass and drag and complexity - less risk, less cost.

The outcome of all this is that the solution to cheap access to space does not lie so much in Technology but in Politics. Equatorial launch will always be cheaper than launch from the continental USA, however, it would be expensive to develop a new construction and launch facility, and all the neccesary technology in a new location.

It is certainly stupid (but historically understandable) to build parts all over the USA and ship them to a launch site.

It may be pragmatic to adopt a non-ideal launch site if all the ancilliary infrastructure exists. (Good roads and frequent large airstrips in the southern USA for example, for multiple fly-back options. Much cheaper than building a dozen military-grade airfields along the length of Indonesia [Though these may exist already, just they may be full of TNI aircraft])

An "ultimately best" solution presupposes huge changes in national and international politics. Perhaps for now we just have to do what we can?

* This leads to a whole new thread - economy of scale is fine, IF you're launching en masse. Unfortunately, we have seen many examples from Titan IV to Ariane 5 of commercial calculations on the number of launches being wildly optimistic, and hence the per-vehicle costs being higher than planned. This is where the Soviet "if it ain't broke, don't replace it" method triumphs. Every new design is another set of sunk costs.

Opportunity and Spirit cost a fortune to design and build as prototypes. Every Mars mission seems to start from scratch and invest in a whole new architecture that needs to be designed and exhaustively tested. I know there are common elements, but both the airbags and parachute for the MER rovers were scaled up beyond their original design parameters and each failed on first test so had to be reworked.

What would be real smart for the future would be to declare the delivery package - the re-entry system and frame to be "finished" (incremental evolution is allowed, but the dimensions and mass are capped). Then, for future missions, you can land anything you like, so long as it fits in that box.

Returning to launch costs, there is no point in designing a new and theoretically more efficient/cheaper launch system, unless you have a commitment to use it at maximum design rate. That requires a political and financial commitment to a long-term space infrastructure mission.

In theory, ISS was such a mission, but IMHO was the wrong choice. Bad orbit (too low, too inclined), and totally wrong philosophy - a zero g station for "research" in space, rather than a practical long-term living experiment with rotation, centrifugal force, and artifical gravity. Such a station would have been an important first step to the planets and asteroids, rather than a junkheap waiting to burn up.

And don't get me started on the Shuttle...

I realise there are no clear answers in all the above, but I think it shows that you need to sort out the ground rules first before we can really address the finnesing issues.