Hi PaQ, yeah the problem is like Pilou shows there, that any particular edge on a surface does not necessarily align in any special way to the isocurve directions.

Here's another example - here is a trimmed plane, you can see the edges are swooping around all over the place:




Yet isocurves from that face come from the "underlying surface" and so are straight lines like so:




You can see there that in the general case there is not really any relation from an edge to one particular isocurve direction, it's only in a special case of an "untrimmed surface" (or an actual isocurve trim) that a selected edge actually maps to one particular isocurve direction.

Something like an "edge slide" I guess would be a somewhat different operation than isocurve generation - maybe something more like offsetting a curve on a surface.

- Michael