22 Jan
2009
22 Jan
'09
3:36 p.m.
On Thu, Jan 22, 2009 at 8:48 AM, Grzegorz Slodkowicz
What is the reason for vertex/edge removal functions in subgraphs being unimplemented? Is it simple oversight / lack of demand or would that pose some problems?
Probably because It can't be safely or efficiently implemented in a generic fashion. There's a tendency for the entire library to rely on vertex and edge indices as opaque references, and removing a vertex or edge can cause a kind of cascading renumbering of elements that is roughly O(V) or O(E) * N (the number of subgraphs). Andrew Sutton andrew.n.sutton@gmail.com