1. When going through the effort of adding catenary, I would make it live. Otherwise I'm not sure how it would help your operational aspects.
2. It's been years since we had a layout with it, but I think it was Fleishmann which fit in under the track and then went overhead which held the proper height.
3. Turnouts and crossovers require some gaping as I recall.
4. It's hard to detail around the track once the catenary is up, so do that first.
As a Nx2 scale (1:80) operator with puzzle palace specialwork used by my numerous cat motors and EMU trains, I prefer the opposite tack. Like that of most of my Japanese-prototype brethren, my catenary is semi-virtual. The poles, support bridges, substation and other support structures are modeled. The actual catenary is virtual - no cat's cradle of wires cluttering the air over the cluster of double slips, and no need for something to hold pantographs down on hidden track (which is far more extensive than the visible track with `catenary.')
There's also the detail that, if you are going to be operating steam or diesel over the same tracks, it's nice for everything to be controlled by the same system. My visible catenary runs from a pair of tunnel portals to an engine change point, and some trains (DMU or with diesel-hydraulic locomotives) operate through.
If you are planning a single track scenic loop with few or no turnouts (most unlikely venue for a GG1,) by all means, hang wire. If you're planning to model real GG1 territory, you might be better off if you lose the overhead. The poles and bridges will still convey the `feel.'
Chuck (Modeling Central Japan in September, 1964 - with virtual catenary)