How does that make a difference with regards to how the floors are aligned? The entire inside of the 8 m diameter section would likely be a single pressurized volume, so putting the location of the airlock/docking port in the middle would be just as inefficient for both layouts.
I'm merely suggesting that an interior deck arrangement like Skylab is more efficient and practical, (see image 1) whereas the SLS hab concept wants to lay the floors out lengthwise, like the Bigelow BA 2100 concept (see image 2). I much prefer the Skylab layout.
Skylab’s layout had problems. The wide open space made space sickness worse and it made it easy for the crew to become stranded mid space unable to reach a wall. It also made it hard for sound to be conducted through the station. Hence why the ISS tends towards narrower modules and a single “floor”. i.e. Corridor like.
Not to argue your points re Skylab, but the reason the ISS tends to narrower modules stems from the Shuttle-launched nature of the U.S./European and Japanese modules. They were all volume- and mass-constrained. The layouts have "floors" only by convention; racks were laid out according to a defined pattern in the original "racetrack" module configurations simply to help prevent possible issues with the crew becoming confused about orientation while working and moving from one place to another.
The Russian segments have their own reasons for diameter, not the least of which is payload launch and integration limits of their own.
Shuttle payload capabilities aren't the only reason for diameter limits. It turns out there aren't a lot of facilities for handling larger payloads. You have to make bigger doors, larger vehicles, higher roofs, bigger shake table, bigger thermal vacuum chamber (most relevant for testing fairings, the current one we use for this stuff can handle Skylab), and to transport the payloads means we need to reroute around bridges, need to move power lines, put on a customized aircraft (these are very rare for bigger than 5m, the one we used for Skylab I don't think is available any longer, and Skylab was just 6.6m).
It is true we might make them a little bigger than 4.5m outer diameter if we built them for launch on EELV or Titan IV or some HLV, but not much bigger. The optimal is not usually all the way over to as big as is possible.
And while it may make sense for a LEO station or maybe even a gateway to use large modules, for a Deep space exploration hab, you want it small and very lightweight because you're pushing it through a lot of delta-v, and it causes the propulsion system to have to be bigger.
A good concept may be to use even smaller modules... A node surrounded by small modules, the number of modules depending on crew size and mission duration. For short missions, no modules. For long missions, use 1, 2, 3, or 4 Advanced Cygnuses docked around the Node (don't know if berthing would work or not... what's the minimum arm size to make berthing work?). The Advanced Cygnuses have 27m^3 of pressurized volume each (7.5m^3 more than Orion), compared to around 75m^3 pressurized for Node 4, so 3 of them would more than triple the DSH's pressurized volume. Advanced Cygnus launches next year, and with 4 of them combined (again, hand-waving the difficulty of tying the buses together), would produce enough power for the DSH to not need any solar arrays of its own, further cutting costs. And we know how expensive these guys are (not very, we're throwing away two of them every year just for unmanned logistics, so less than $250m including launch), and they're already produced en masse for ISS CRS. They'd need more MMOD shielding, but that's not a huge expense. But we could build them only as needed. And the big advantage is economies of scale: We are likely to only produce one or maybe two of these DSH for decades at a time (assuming we're reusing them), while we could produce a dozen or so of these vehicles for both living area and logistics, and we're already making 9 of them (including the regular Cygnuses, including the COTS Demo 2), so we'd already be benefiting from economies of scale vs starting from scratch with a big module.
The advantage of using ISS heritage is: cost, cost, cost. And speed. Could build it ASAP. Could maybe even afford it, even in tight budgetary times, while I'm not sure we could afford building a new Space Station essentially from scratch (that used non-standard and thus much more expensive ground infrastructure) while paying for both SLS and Orion. ISS heritage is the way to go, IMHO. We've got a decade or more of experience, the development is paid for, we even have spare parts which could save a significant amount of money in fabrication, let alone the time advantage. For wherever possible, ISS heritage makes a ton of sense.