[graph theory] to cut the edge of the time should be used to do clues or use the edge to do clues?

When looking for cutting edges, if you use points to do clues, such as a to B there are two without the edge.

Now there is a case of heavy edge, if now from point A to point B, according to clues, we think a is the father of B, then we go back from point B to the side must be gone. In this case, if low (b) > Dfn (a), then our algorithm will consider a cut edge between A and B. This is obviously not true, and we should have a reasonable way to solve the problem.

We use the side to do the clue, or just the hypothesis. From A to B, we write down this non-edge, then B can no longer from this non-side back to a point. But B can still return from another (that is, the heavy edge) to a point, in fact, a "father" to remove the special attributes, in any case, as long as B has a side can go backwards, are considered to be atavistic side. When B returns to a, it is bound to update the current low (b), so finally Low (b) =DFN (a), it can be judged that the edge between AB is not cut edge.

In fact, compared with the implementation of these two algorithms, we can know: by cutting edge can be cut point, but through the cutting point may not know cutting edge.

[graph theory] to cut the edge of the time should be used to do clues or use the edge to do clues?