Well, you are talking about going back to 1972 or so and getting a very different Shuttle configuration! That would probably have been very good to do.
The Saturn Shuttle was a good idea I think. We have to remember that what everyone was dreaming of, what the majority of the aerospace community thought was both feasible and clearly economical, was a single stage or two stage system with both parts fully reusable. Most agreed two stages would be needed and that the first stage would be piloted and would fly back to the launch site as a jet-propelled suborbital spaceplane, while after separating from that a second stage orbital speed spaceplane would contain all the propellant it would need to reach orbit, and deploy cargo from a cargo bay, take on down mass and then return for a horizontal airplane like landing, as the first stage had done. This is what everyone was thinking of. And later I thought they were insane to want an orbital spaceplane that massed far more than the cargo, because it meant putting a lot more mass into orbit than they had to. But I realized that of course airplanes do this all the time, and that since propellant is relatively cheap compared to the cost of making a rocket's dry mass, it is OK to use 5 or 10 times as much propellant if you are able to use the spaceship over and over dozens or hundreds of time with only modest refurbishment. When you look at it that way, the real historical STS was not as crazy as it can be made to look. The reason it was expensive was not just that it was a Saturn V sized monster delivering a Saturn 1B sized payload--mainly it was that refurbishment costs, including time needed to accomplish them, were quite large and not modest. I've never found a source that pins down just how much it realistically did cost because the program is very political. NASA had ways of playing fast and loose to lowball the cost and hide it with subsidy; latter-day Monday morning quarterbacks who want to pour on as much bile as possible I suspect of greatly inflating it, by such tricks as smuggling in the development cost into each operational launch. If the Shuttle had been made by a private firm that had taken out a great big loan to develop and build it and all the launch facilities and so forth, then they certainly would have to include the cost of recovering that investment so they could pay back the lender with interest, so amortizing development would be a line item in each launch. However--the Shuttle development was paid for the US taxpayer in the 1970s, it was water under the bridge and NASA had no obligation to collect the value and pay it back to the US Treasury. So it is wrong to include those costs in the operational cost of Shuttle missions, though it is certainly an instructive statistic to compute to get an overall sense of the economics of the system.
The ultimate TAOS configuration adopted was meant to reduce the development cost of the Shuttle system to a low level that the Nixon Office of Management and Budget was willing to accept. To try to develop the "real" configuration with two liquid fueled flyback stages was found to be considerably higher than OMB would tolerate. It was believed that the solid boosters would be much cheaper to develop (true, at least with a clean sheet booster to compare to) and yet would be reasonably economical to reuse--quite false. Again no one will give me exact figures, but it made little sense to try to recover and refurbish solid fuel engine elements. Actually maybe it would make sense to reuse the nose cap section and the nozzle section, since these would be simple "plug and play" items, assuming that the hot exhaust gas they contained and guided did not erode or otherwise wear them badly. It is not clear to me these elements were in fact reused. It is clear that the silliest element to attempt to reuse was where most or all the effort went--the solid fuel segments! These were little more than lengths of high pressure containing wrapping to hold the solid fuel in place as it burned, and would be quite a small part of the cost of the segments, which mainly would be the formed solid fuel grain itself. Clearly in retrospect, either all attention to recovery should have gone just to the nozzle, or maybe the nozzle and nose pieces, but the fuel segments should have been regarded as disposable. Add to the cost of the quixotic effort to "reuse" these simple cylinders by shipping them back to Utah the cost of recovering the whole spent solid at sea, and I can believe the claims that the added costs were about equal to the savings, and many suggest they were more. But it was very important to the political selling of the TAOS Shuttle that the solids would be reused, and optimistic projections of savings helped sell the idea that using solid booster elements would be cheaper than developing the Saturn V first stage into a suitable liquid booster.
The Saturn Shuttle idea had several variations too. One version famously shown as a model for President Nixon's desk was an apparently unmodified first stage on which an upper oxygen-hydrogen external tank of the same diameter was mounted, with the Orbiter looking much as the historic one stuck sidesaddle there. In this version, the first stage would simply be expended. That made it a tough sell versus allegedly reusable solids! Other variations proposed putting some wings on the first stage and presumably other fancy developments, and a crew cabin (it was assumed all fly back stages would be crewed, and the astronauts would not have it any other way insofar as their opinion counted, and it did count). It would thus have jet engines and would have to take rather substantial thermal stresses to survive undamaged--it would not be going fast compared to the upper stage but it would be going very fast compared to standard issue airplanes, even military supersonic fighters and bombers, and it would get very hot. So it would not do to just weld some wings onto a standard aluminum S-1C first stage and call it good! The whole thing would have to be redesigned, made out of titanium or high temperature steel, the added weight would make the size of the upper stack take a hit--it would be a big project. Of course out of it we'd get a truly reusable first stage, although of course there was no way to predict whether it would turn out as with the historic Orbiter that they underestimated the refurbishment costs. I've read a good alternate history timeline that suggests that for a similar amount of budget money they might have decided to go ahead and invest in making this first stage, and postponed and cheapened the upper stage Shuttle by making it ride on an expendable second stage basically a version of the Saturn S-IV, the almost common second stage of the Saturn 1B and third of Saturn V--the investment there would be to cheapen production while stretching it a bit. This would make the actual Shuttle Orbiter a much smaller vehicle with no main engines, only orbital maneuvering system, and a small payload, the idea being that bigger payloads could ride without an Orbiter on the second stage. A key idea of this alternate history timeline was that a portion of the first stage propellant would be reserved to brake the first stage down to a much lower speed after the second stage separated, which would trade off some theoretical payload capability with greatly reducing the heat dose the structure had to tolerate. This would allow use of more standard, cheaper and lighter aeronautical materials to make the first stage out of, thus cheapening and speeding its development. The timeline is called "Right Side Up." This refers to the idea that the booster is a bigger, heavier system and therefore more costly than upper stages and so ought to be the first place someone looking for a reusable rocket system should concentrate. Instead the historical TAOS shuttle design concentrated on recovering an Orbiter that would also include the main engines, in effect being preoccupied with recovering the upper stage, and taking a cheap to develop shortcut on the massive booster system.
I think I would advocate for a third approach to the recoverable liquid booster--splash and drag back! Even expendable booster designs often have their first stages splash down or crash in remarkably recognizable condition. The idea here is to use more structural weight, using steel instead of aluminum, to make a robust stage structure that can take the heat of reentry and simply splash down in the ocean, to be retrieved by a ship that just drags it back to port at the launch site, then inspect it for damage, test and refurbish as needed, hopefully minimally, and then if it checks out or is patched up suitably to then check out, put it back in the VAB and add the upper stages and payload to it, haul it out (being all liquid fueled, it should be much lighter than the historic TAOS shuttles with 1200 tonnes of monstrous solid boosters loading down the crawler--a common feature of all liquid boosted systems of course!) and launch it again. I even wonder if making so strong that the propellants can be pressurized to the chamber pressure of an F-1A engine, and thus the mechanical turbopumps and gas generator systems of that engine can be eliminated in favor of simple pressure feed would be about the right strength so it can stoically endure repeated launches without accumulating serious damage. Such a pressure fed system would also greatly simplify as well as lighten the engines and thus speed up refurbishment considerably and save costs there too. The simple burn to full speed then aerobrake and splash down system also gets the maximum use out of the propellants for the upper stack and hence payload. It is still necessary to redesign the stage, but it is a relatively simple design; the use of very strong structural elements presumably simplifies many things. It might seem fatal to raise the dry mass of the first stage, but actually the burn out mass to be considered includes the full wet weight of the whole upper stack; as long as the dry mass of the first stage is small compared to that, it makes little difference to performance. Pushing the dry mass up to a significant fraction, but still much smaller, than the upper stack gives margin to make it very very strong indeed! So massive would it be I doubt parachutes would do much to slow it and it would have a very high subsonic terminal velocity, so I figure retrorockets of some kind would be needed to slow it down to acceptable splash speeds, but doing the math on it I don't think that is a big problem.
A common feature of all Saturn Shuttles was that the layout did not allow for the possibility of ground-lighting the upper stage hydrogen engines. And that is a good thing for a more practical Shuttle program, since a huge part of the cost of developing the historical Shuttle was developing the hydrogen-burning SSMEs, which had to achieve decent Isp and thrust at sea level despite being hydrogen burners. This meant that the reaction chamber pressures had to be raised to 200 atmospheres to give hydrogen-oxygen exhaust a decent performance at sea level pressures, which led to all sorts of extra complications. All this also had to be reusable hundreds of times, which made the project even more insanely massive. Whereas, if it is accepted that the hydrogen engines will not light until the rocket is at a high altitude with much thinner air, already developed engines like the J-2S from Apollo's Saturn upper stages could be used instead. This was not a popular option unfortunately; the J-2S was in my opinion an excellent engine but its manufacturer was quite bored with it and would not have enjoyed a contract to turn them out in quantity; they wanted the challenge and high payments for developing the advanced hydrogen engine to be used at sea level. In fact the other "classic" shuttle concepts that called for two spaceplanes would not have used the kerosene-oxygen mix the Saturn Shuttles would settle for as an Apollo legacy; they wanted these same sorts of sea level high performance hydrogen engines for the booster as well as for high altitude performance on the orbiter, and cheerfully embraced the bloated volumes needed to provide hydrogen at its very low density in quantity. The idea of continuing to use the F-1 engines in their improved F-1A form was unpopular, as was using the similarly upgraded J-2S. Everyone wanted new toys, not to frugally reuse the legacy equipment they had from Apollo. They wanted a new Apollo in fact, and starting over from scratch was the preferred approach.
So--I think we would have done far better to develop a Saturn Shuttle. Mind, we'd still get into trouble with the ceramic thermal protection tiles, which by their nature are fragile and need to be laid out on the hull in very great numbers, making a very labor-intensive and slow turnaround process. There were alternatives, but they had their own drawbacks. I think ablative would have been quite unacceptable for instance--although I do wonder how workable it might have been to put an ablative layer atop the ceramic, so that a relatively thin layer of ablative might have taken the brunt of the highest entry heats, then burn off completely and smoothly exposing ceramic, or maybe metal that would be good enough once the ablative had gotten past extremes, that would char off all the ablative and thus present a predictable aerodynamic smooth surface for the glide phase of entry. Obviously this would be a labor intensive costly delay in refurbishment, painting or plastering or gluing on another layer atop the fixed permanent lower temperature TPS before each mission, but in retrospect hindsight tells us it would be cheaper and faster than managing the tiles, and also safer.
You also seem to suggest that you'd favor a fourth kind of Saturn Shuttle, one that does not use wings and does not splash down, but rather uses reserve rocket fuel to retro-brake the first stage, shove it back to the launch site, and have it rocket-land on its tail at the launch site like a Falcon can. But I don't think that is realistic for a 1970s program! One reason Falcon rockets can land their first stages like that is that the standard Falcon has 9 Merlin engines; it is easy then to get low thrusts by lighting just a few of them and throttling. Variations on the Saturn V might possibly have gone up to 7 engines, a filled hexagon, I suppose, but it would not work well I fear to light 2 or 3 F-1A engines for a controlled gentle landing! Also, the option of flying the Falcon first stages back is pretty costly in terms of payload that can be launched; it is nice they have the intermediate one of landing on a drone ship downrange. Needless to say the option of a robot drone ship in 1971 would have been more far fetched, as would the confidence that a robot controlled system could successfully manage a rocket landing at all. So, overall that might have been a clean sheet option in the later 2000s, but never for the original Shuttle.
So I do support the idea that a Saturn Shuttle would have been the smart way to go in 1971. But the scenario I am working on assumes that as in real history, the Shuttle Decision went much as history and we are stuck with the solids made by low bidder Thiokol which fell short of promises, especially in reusability, the nightmare tiles on the Orbiter, and the long and extended process of expensively developing the SSMEs. All of this is water under the bridge and NASA is committed to the STS system for the 1980s. Then Challenger is lost. I would like to explore the possibility of developing more flexible and more cost-effective use of the expensive Shuttle component elements. In fact this scenario here is focused even later than the Challenger post-mortem, more in the Columbia post mortem period of the mid-2000s. Here, the recognition that the old Orbiter fleet is going to be retired soon forces consideration about what to do afterward. Historically, the Constellation system was proposed, and when one digs into that decision, I see a couple things I think are pretty ugly that led to grossly unsatisfactory results. For one thing the NASA leadership was not interested in retaining the capabilities that the Orbiter did give us, such as the ability to put 7 or 8 people into low Earth orbit and support their doing useful work there. With the development of the ISS of course it was no longer necessary to have a giant Space Winnebago to hold a miniature SpaceLab in, but neither was NASA's leadership interested in making a smaller space bus to take such large crews to ISS and do their scientific work there. They wanted to shut down ISS within just years of getting it operational, and instead focus only on long distance missions, to the Moon or Mars. This mission is all Constellation was designed for, it was not meant to assist in operating the ISS which the administration proposed to abandon! Thus we would have no economical means of reach low Earth orbit with large crews, or even crews of any size at all, only the expensive option of wasting an Orion mission in a spaceship oversized and overpriced for LEO work. Secondly, despite the evident desire to make a clean break with the Shuttle legacy, they were under some pressure to develop "shuttle derived" systems. The idea here was not so much that Shuttle systems had in fact solved some problems well and we should continue to use that successful aspect of STS. No, it was that on Earth, established contractors such as ATK Orbital (heir of Thiokol) wished to continue to see their facilities used to supply NASA, never mind whether their products were the technical best way or not. And NASA being subject to political pressures, such as the powerful lobbying these established Big Old Space contractors could wield, meant they had to think inside a box that would reuse STS tech whether it was a good fit to the problem at hand or not.
These forces shaped the Constellation proposal. For crew launch, we were to switch over to the Ares 1, which would put a hydrogen fueled upper stage on a solid fuel stick. Since Orion was designed for deep space missions and added some features such as a radiation shelter that Apollo lacked, it was grossly overweight for a mere LEO mission--but had Constellation gone forward, we would have shut down ISS and had no LEO missions anyway. Thus it was too big to be boosted into orbit with a standard issue Shuttle 4 solid fuel element stack; ATK Orbital would be forced to develop a 5 segment stack instead. But the old Thiokol works would go on churning out more sold segments! That was the important thing you see. Then, big spacecraft such as the Moon lander and a TLI stage to send an Orion/lander combination to the Moon or more distant places would go up on an Ares V--this was presented as cheap to develop because it was suggested to be a Shuttle style external fuel tank, with SSMEs on the bottom meant to be used once and thrown away, the payload on top, and familiar side mounted solid boosters. The catch, from the point of view of claiming economy by using already developed Shuttle parts, was that the new tank first of all would suffer new stress patterns with its payload on top and its engines on the bottom, instead of those things being sidesaddle and together as on the Shuttle. And second it was to be stretched to be bigger than the old expendable tanks. Being more massive again the solids from the Shuttle would not cut it; if we were lucky the same 5 segment stack used for Ares 1 could do but there was a desire for a 7 segment stack. Thus every allegedly "Shuttle Derived" element would really be a new development project, nothing was taken as given by the Shuttle program. The contractors would receive large new payments for development on top of continued demand for their existing facilities to be used for components.
I would be less outraged by all this if in fact they had delivered with a useful product in time for the retirement of the Shuttle. But this has not happened. Obama might be blamed for refusing to continue Constellation but Congress forced the Ares V program to go forward anyway under a new name but with the same pork-maximizing, results minimizing priorities in place. SLS is too expensive to be used more than once every other year, and leaves no budget for something to ride on it, and has not been tested.
So I was wondering, what sort of spacecraft might we make with genuine Shuttle Derived hardware? I believe for instance that an adequate Earth to LEO crew vehicle could be based on a booster made from a single segment of the Shuttle SRBs! That would allow for something like 10-12 tonnes in LEO which ought to be good enough for a taxi to the ISS for 5 or more crew. Two such single-segment boosters should be able to put cargoes as massive as the Shuttle could up, flanking a central liquid hydrogen fueled upper stage. A whole family of possible booster-tank combinations could be studied, including those that would be big enough to use up one or more SSME to dispose of it. The combinations with the most economical features could be chosen for an alternative kind of Constellation, one that preserves our LEO capabilities while being extendable to bigger configurations with massive capacity. The old space Mafia could be gratified with ongoing indefinite demand for their products while achieving much greater economies per ton to orbit.
By the mid 2010s of course SpaceX was showing that a clean sheet approach was more economical still. But I am talking about an Alternative Constellation, one that might have gotten operational, at least with some elements, enough to preserve our LEO capabilities, before STS was finally shut down. Meanwhile to a great extent the new system uses the old elements supported by the Shuttle Mafia, who are gratified. Versus actual history, small rockets just to ISS would be using just a trickle of ATK's solid segments, but American capability to reach ISS would never be lost, and the line item cost of these launches would be far lower than a Shuttle mission. Perhaps with the pace of Shuttle launches coming down due to the alternative small crewed vehicles flying to ISS instead, the Shuttle Orbiters could be stretched to the present day or beyond due to having fewer missions to fly, justifying postponing their retirement.
That is the scenario then. We stick with legacy Shuttle tech because the suppliers have political influence, and it is still a bit early to have trust in SpaceX's ability to get it done cheaper a different way. American crewed missions continue to launch from the USA and we have the newer Orbiters handy to carry out special missions while routine stuff is more and more done with the Shuttle derived but not Shuttle vehicles carrying them.
It is not better than what we might do now with SpaceX developed options. But it is better I think than what did happen between 2006 and today, that is the last 12 years!