Hello Bernhard

The opposing curves in the image you supplied are exactly the ones you noted, same case would be the two rail ones.
Remember at the end, Nurbs surfaces are just four-sided patches so there´s not much point at which you can get lost with them.

As per your question regarding spans, it would mean it needs more points, yes. Since spans and editpoints actually converge in the same point in space you can pretty much predict where will the span fall by watching at Edit Points. To change the spans the best way of doing so would be, going into the Attribute Editor, noting the spans of the curves, go into Rebuild Curve and rebuild it keeping the ends, but specifying the new set of spans. That should take care of it quickly and without mess.

Direction can be noted in Nurbs curves by watching the litte U-icon in the curve and it´s relationship to the little cube one. Direction of the curve will go straight from the cube to the U to the end of the curve. It´s called U direction in curves.

On a little side note @Mike: Why should he take CVs into account for spans, since they don´t match up, and can be subject to confussion.
I mean it because if you didn´t have a way of checking the spans in the Attribute Editor, I could pretty easily trick you into thinking there are more or less spans by looking at CVs. -CV view does not show what degree the curve is, only the CVs, thus to know spans by looking at CVs you´d need the degree info of the curve right?-
That´s why I still believe he could look at EP info to know how many spans there are. (although truth to be told I would check in the Attribute Editor just like you)