I did my experiment. I had one surface that had one naked edge. (All of the surfaces in this model had been 'Joined' and that was not a problem -- I then had one Joined object. 'Details' nevertheless indicated 15 naked edges. So 'Join' did not fix the edges.) This surface looked perfect and zooming in to the limit did not reveal any error. So I completely regenerated this surface by creating cross-sections (with 'Iso'). I moved all the points (26) 'On' to the edges of the adjoining surfaces. Only in the case of one of these points could I see (very close zoom) that it was obviously not on the edge. (However, I've found that any ambiguity when indicating an endpoint, intersection, point, etc. is suspect. After resetting the endpoints, as I typically do, this ambiguity disappears.) I then used 'Network' to make the new surface. Again, it looked perfect. And again, the same edge was shown as Naked. From this I conclude that being aware of "Naked edges" is not useful information, as there is nothing that we can do to correct the errors -- at least if 'Network' is used to make the surface. You point out that two surfaces may have different numbers of points at a common edge. Certainly this is unavoidable. I am given to conclude that this is the cause of a "Naked edge" and if this prevents a joined object from being regarded as a "Solid", then we will have to accept that this is how it goes. If 'Join' results in one joined object, that's as good as it gets.
|
