I've really hesitated to get involved in this thread, not only because the set-up and use of hollow planes is somewhat thorny and about which there is no general agreement, but also because I'm not sure anything I have to say will be all that useful. But, I've finally decided to take a stab, because I'm not sure this thread, as it stands now, has been very helpful either.
From the outset, some 34 years ago, I've found hollows to be somewhat problematic. And, I still think there is no simple "magic bullet" approach which will satisfy everyone. Or, maybe, even the majority. But, after working with H&R's's almost daily, especially for the past six years, I've come to the conclusion that the approach which generally affords the best outcome in the widest array of usages is one which tries to keep the full profile of the iron as closely matched to the profile of the sole of the plane as possible.
I say this because it allows the plane, whether set for a heavy or fine cut, to cut the whole profile. This allows the arc of a convex portion of a moulding to be carried right up to a fillet, say of an ovolo, or the inflection point of a cyma curve. And, yes, this does involve the risk of a tendency to leave tracks at the margin of the cut. This can usually be minimized, though (assuming your hollow fits the arc you're trying to create), by adjusting the iron for a very light cut for a finishing pass or two. (I've just pulled out some sample mouldings I've done like this to double check myself, and the resulting "tracks" are only visible if held up to glancing light and would readily disappear with a light sanding.)
In some situations, as Al suggests, a hollow with a slightly larger radius can be used to ease these tracks. But, this isn't always possible. For example, if one is creating an ogee based on a sixty degree arc, the larger plane won't fit in the necessary space and will definitely foul the concave portion of the moulding if it needs to be used anywhere near the inflection point.
Sometimes, despite one's best effort to match the profile, the hollow plane is prone to leaving a heavier track on one side. In this case. it is sometimes possible to make succeeding passes with the plane which "walks" that track toward a fillet, where a light cut with a rabbet plane can eliminate it without crating a distractingly noticeable flat. And, to return to the ogee, or reverse ogee, the most difficult area is the inflection point. My strategy, there, is to get the concave portion very close, but intentionally leave material for a shaving or two near the inflection point, then work the convex portion. In the event, actually the likelihood, of leaving a track at that point with the hollow, I can then remove it with the last pass or two with the round plane.
As to the types of profiles which hollows can be used for, assuming the radius of the curve is appropriate to the plane, they can be used for arcs of any number of degrees - assuming there is room for the plane to be used. This is accomplished by changing the orientation of the body of the plane to position the sole so that the iron can cut where it needs to. This includes astragals. In fact, I'm puzzled how a side round (with a convex sole) could be used to help create an astragal (also a convex shape).
And, since I've mentioned sanding, I guess it might be useful to get into that a tiny bit. If one is sanding the other show/visible surfaces of a piece of furniture (which is not "cheating"), then a light sanding of the mouldings will be in order as well. If done with care, possibly including specially made sanding blocks, the details of mouldings can be preserved while doing this.
At Rochester, as Tom pointed out, we also touched on scraping - because someone asked in this context. As I think I said there, I have done that and it can be useful, but I really like to try to keep my planes tuned so as to be able to keep scraping and sanding to a bare minimum.
I have no idea if this has been helpful or not.
Don McConnell
Eureka Springs, AR