When we get into the more utilitarian phase of space development, then I think you would want something like shipyards where there really is a big micrometeorite shield (yes, as a sphere) with the inside filled with scaffolding. Robots scoot around do work with old and new hardware.
If you think about nesting ROTATING structure inside of other pressure-envelope structures, then you're getting into some really crazy stuff. Are there designs that might make sense? Maybe, I don't know, I guess I wrote a blog about it
Windows are at a huge premium on cruise ships. It'll be much worse on space ships. But it's possible that inside windows will eventually look out onto something more interesting than the black void of space, so an inside window may be preferable. One of the reasons we look out the window in a car or on a boat is to establish the horizon and fight motion sickness. If you look 'outside' of a rotating space station - especially a rotating space station orbiting a planet or moon - you'll head rapidly in the exact opposite direction.
I think my train of thought makes a bit more sense for moon and asteroid bases rather than free-floating orbital structures. And asteroid bases - on the right asteroids - are probably going to house most of the people.
The ISS design definitely doesn't scale up. Everything off axis is a lever arm and the bigger you make it the more the whole thing tries to twist itself apart.
For modularity it might make more sense to use nesting. A building inside a building has no seams. Doors only need footpaths between them, not hard structures. The inner building can be used for shelter in case of an accident, and can be run at higher pressures than it could in hard vacuum.
In the tinker toy model you would tend to have to keep repurposing buildings because while the size may be appropriate, the older structures may get pushed farther and farther from the center of the action, rather than staying in the center of the action.