In light of the latest assembly mission to the ISS, Clark Lindsey points out one of the many absurdities with the ESAS approach:

I’ve noted before that I find it odd that a fundamental goal of the Constellation project design is to minimize in-space assembly. This is a task in which NASA has actually become quite good. If NASA went to the next stage and combined its in-space operations capabilities with fuel depots and a space tug, it would have the tools and skill sets to do some amazing stuff, especially if it worked in close cooperation with private ventures like Bigelow.

Unfortunately, as with building Ares 1/5 instead of using existing (e.g. EELV) or nearly ready vehicles (e.g. Falcon 9, K-1), the aim seems to be to time warp back to 1972 and continue on to Apollo II, ignoring much of what was learned and developed subsequently.

Indeed.