One of the problems with the Shuttle was that it was an enormous rocket with huge fixed costs. After the retirement of Energia/Buran, its Soviet copy, it was, by a huge margin, the biggest rocket in the world. The fixed costs of operating it was about $2 billion a year even in years the Shuttle did not fly at all.
It did one thing very well: carrying crew and cargo to a space station and helping to build it.
It was also capable of servicing, repairing and upgrading the Hubble Space Telescope. Arguably, it would have been less expensive to plan from the beginning to launch several progressively improved versions of the telescope.
It could and did recover or repair other satellites for reuse, but those missions were much more costly than launching a replacement. It could and did put satellites into low earth orbit, but it was found that other rockets could do the same job for less money and without risking a human crew.
In the end, it was a special purpose vehicle, used to support ISS and occasionally service Hubble. NASA was essentially the only user, and so absorbed the fixed and variable costs year after year, severely restricting NASA's ability to fund other human spaceflight projects. NASA was unable to simultaneously operate the shuttle and develop a replacement.
Other, smaller rockets are more flexible. Atlas, Delta and Falcon can split their much smaller fixed costs among several government agencies, and possibly commercial customers as well.
In retrospect, building a huge specialized launcher with NASA as the only customer was a mistake if there is a reasonable alternative
Unfortunately, there are many in Congress and at NASA who want to repeat the mistake, building a huge, specialized launcher with NASA as the only customer, reusing as much Shuttle technology and infrastructure as possible.
The attractions of this are obvious if your district is making Shuttle components or operating Shuttle infrastructure, or if you work at a NASA center that expects to oversee or operate the Shuttle-derived launcher.
For everybody else, this is a bad idea. At the funding that NASA can realistically expect, a specialized mega-rocket is not the best way to get beyond low earth orbit.
We can return to the moon or visit asteroids with existing launchers, using spacecraft assembled and loaded with propellant in low earth orbit. Ed Kyle explains.
At worst, we use a brute force approach and storable propellant for the departure stage. This is less efficient, pound for pound, then liquid hydrogen and hydrogen, but it's proven technology. We know that storable propellant can be stored for long periods and is routinely transferred in low earth orbit.
I believe that we can do better. It should be possible to store cryogenic propellant with limited boiloff in LEO and transfer it in orbit. This would significantly reduce the mass required in LEO for any given mission. If not, we can fall back to storable propellants.
This approach has several virtues. NASA would avoid the need to develop and maintain their own unique launch system, freeing more money for spacecraft, other systems, and missions. At the same time, shipping propellant and hardware to LEO would be a significant new market for which currently underutilized US launch companies could compete, spreading their fixed costs over more missions and significantly lowering average launch costs.