As I have argued before in the pages of this journal, NASA has over its history employed two distinct modes of operation which, for shorthand, we can call the “Apollo Mode” and the “Shuttle Mode.” The Apollo Mode, which prevailed in the human spaceflight program during the period from 1961 to 1973, involved first choosing a mission goal, then developing a plan to achieve that goal, and then designing and developing the hardware and technologies needed to implement the plan. The hardware set is then built, after which the mission is flown.
Where to begin? NASA, over its history, has employed two distinct modes of operation. The Apollo Mode began in 1961 and reached its peak in 1969. After our first manned landing in 1969 the program rapidly lost momentum and Saturn V production ended in early 1970. The last lunar mission was in 1972. During the Apollo Mode, we chose to use manned space flight to demonstrate our superiority over a rival superpower. We devoted enormous resources to that goal, and once we achieved it we lost interest. The Apollo era was a brief spasm when national pride and prestige inspired massive spending to meet an arbitrary deadline to win bragging rights over the Soviets.
The other mode was everything else: NASA before 1961 and after Apollo: in short, the vast majority of its history.
Also, the Apollo program did not simply choose a mission goal and then develop the technology to achieve that goal. The F-1 rocket engine that sent Apollo rockets to the moon was developed to meet a 1955 USAF requirement. Requirements for what would become the Saturn family were issued in late 1956 or early 1957. We didn't settle on lunar orbit rendezvous as our plan to reach the moon until 1962.
The Apollo program did not follow the Zubrin ideal of choosing a single goal and then building the technology to achieve it. Which is fortunate, because if we had the technology wouldn't have been ready when we wanted to go.
He then follows up with:
The potential utility of orbital propellant depots — basically gas stations in space — as a way to enable manned missions to the Moon, near-Earth asteroids, or Mars has never been established.
It hasn't been established because it hasn't been attempted yet, like lunar orbit rendezvous in 1962. It's clear that it has the potential to be tremendous game changing technology, allowing manned missions beyond earth orbit to be accomplished either with existing launchers, or with much smaller new launchers than we'd need without it.
Note the careful phrasing of Zubrin's complaint: orbital propellant transfer is done routinely by the Russians to reboost ISS, and has been demonstrated by the U.S. Orbital Express mission. What hasn't been demonstrated is the orbital transfer of more efficient cryogenic fuel. If it can be demonstrated effectively, it has the potential to make plans that depended on much larger launchers without propellant transfer obsolete.
Zubrin's pet Mars architecture is one of those plans, which may explain his hostility to propellant depots