Well, to build a station like Skylab, you need a big rocket to launch it all at once (or in just two or three big lumps if you want a bigger station). And the US just didn't have a big rocket. So the thing would have to be cobbled together from modules no matter what. More like Mir than Skylab. Not that there would be a major difference between those, though.As for whether one needs a truss structure for the solar panels... Well, one needs a shitload of panels to power up the station (or then a fission reactor, but those are a strict no-no), and the ways of fitting them onto the modules directly are pretty limited. Mir was horribly user-unfriendly in that respect. Of course, Skylab was even worse, with nonexistent pointing capability - although as a solar observatory, the structure was ideal.
That said, there would be a zillion ways to build a better modular station than ISS or Mir, the main improvement being greater standardization so that the modules could be arranged into compact and structurally strong clusters yet easily changed when needed. The best way would probably still be to do it like the basic Mir concept, and like NASA wanted in the 70s - a central habitat hub and changeable lab spokes. That would give the shuttle something meaningful to do, too, namely changing the modules.
Still, I think ISS is an entry requirement for manned space travel. So it is dangerous and man-intensive to build - but most future projects will be more so. If we can't build this turkey, then we can't do most of the worthwhile projects, either. Things are only going to get harder from here on.
If people actually came to believe space exploitation was safe and sensible and cost-effective, thanks to *one* hyper-engineered silver-bullet space structure that actually had those qualities, and then lost that faith when not all the bullets could be made of silver.. that would spell disaster for space travel. There are too many false expectations already.
Timo Saloniemi