The ISS has operated on a fairly consistent model for orbital assembly that requires three major elements in order to function:

1.                   A minimum 20,000kg lift capacity, payload diameter of 4.15m in the launch system(including both Shuttle and Proton launches)

2.                   A Robotic Manipulator System (RMS) for berthing modules

3.                   Extravehicular Activity (EVA) or spacewalking for completing connections

Russian space station assembly (including ISS, Mir, and Star Modules for Salyut 7) have used automatic docking rather than RMS berthing.  One should not downplay EVAs role in Russian space station assembly; as early as Salyut 7, cosmonauts were installing additional panels to her solar wings by spacewalking. 

The ISS has proven that this model works for both logistics and assembly, although I think using a shuttle integrated logistics module that can be unloaded from within the Shuttles bay (Spacehab and Spacelab) is more practical than modules that need to be berthed to the station using RMS before loading and unloading (MPLM; Multi-Purpose Logistics Module.  Before examining the station assembly manifest, I thought that the MPLM was left on the station between shuttle trips, as its design seemed to favor that purpose.)  By using a fully reusable launch system with low hardware costs, it should be possible to make this model much more cost effective than Option C.  The bottom line is, that the ISS is actually too forward-looking and futuristic for current launch vehicle technology.

Orbital Servicing

If our first fully reusable launch vehicle is much smaller than expendables like Ariane 5, Titan IV, and Proton, there may be a market for servicing satellites, as the cost for servicing them might dip significantly below the cost of replacing them.  (For GSO satellites, a vehicle operating from a station might be required.)  The Space Shuttle rescued several satellites whose upper stages failed to ignite (a rare problem these days, and those that do fail wind up in orbits unreachable by the shuttle or ones that decay too quickly.)  It might be possible that a future satellite launched at a cost of $300 million might need rescuing by a Bluestar(qv) that costs $5 million to launch.  Through repeated visits to the Hubble Space Telescope, and servicing of other satellites, the Space Shuttle has proven the feasibility of orbital servicing.

Lifting Entry

The Space Shuttle has taught us a lot about the vagaries of atmospheric entry.  The wings went through two structural revisions because aerodynamic force theories before the shuttle flew werent perfectly accurate.  Plasma effects forced revisions on the elevons thermal protection (Columbia remained relatively unmodified in this regard and, as of this writing, may yet prove to be the cause of her demise.)  The Shuttle has proven that lifting entries followed by dead stick landings are relatively easy and safe.   We also have experience from ASSET and X-23.  It should be fairly easy to develop entry lift and entry protection for the next generation of manned spacecraft using our knowledge gained from the Space Shuttle.

Backup Recovery Modes (BRM)

This lesson was taught not by our successes but by our failures.  During STS-33's break-up, the crew cabin survived until it hit the water.  Theoretically, all it needed was a parachute system to soften the landing, and a couple of balloons to keep it afloat. On 6 December 1957, after the first Vanguard orbital attempt exploded, the tiny satellite still worked after falling through t he fireball to the concrete. 

The following quotes from an introduction to a documentary on Challenger rammed backup recovery modes home enough for me to write specifically on them:

KENT SHOCKNEK (KNBC anchor, Los Angeles): If the oxygen(sic) packs were used, that means they survived the explosion.

JAY BARBREE (MSNBC science reporter): But you can rest assured, they did not survive the impact.

More specifically, about forty minutes into the documentary: 

ROGER BOISJOLY (Former Morton Thiokol Engineer): The astronauts were alive.  The vehicle continued to climb and fell from some nine to twelve miles up.  (Approx 45-60,000 feet.  PAO stated immediately after the explosion, before he realized it, that Challenger was at 9 nm altitude, so I think the apex is more like 65-75,000 feet.)

These conditions are quite probably about the easiest conditions to execute a backup recovery from. 

Backup Recovery Mode, as I call it, is the building into the crew cabin the ability to survive a catastrophic failure of the rest of the vehicle.  Launch escape is escaping the catastrophic disaster, and that's different.  Backup Recovery Modes are also applicable in entry situations (although it likely would not have helped Columbia.)  Because the survival of backup recovery modules is passive, no reaction system is needed.  The challenge lays in packing enough systems into the module and making the module itself tough enough to survive most likely catastrophes.  After surviving the catastrophe, the systems must then recognize the catastrophe and react quickly enough to save the crew.  Here are some ideas for a Backup Recovery Mode for an STS-like vehicle: