On Thu, Feb 25, 2010 at 7:24 AM, Minh Nguyen <nguyenmi...@gmail.com> wrote: > Hi Martin, > > On Thu, Feb 25, 2010 at 6:45 AM, Martin Rubey > <martin.ru...@math.uni-hannover.de> wrote: > > <SNIP> > >> At least from one point of view it makes sense to make such a choice: >> when you do, the incidence matrix times it's transpose gives you the >> Laplace matrix of the graph. > > I was confused by the interface name incidence_matrix(). For an > undirected graph object G, I would expect G.incidence_matrix() to > return the unoriented incidence matrix of G. Imagine my surprise when > the result was actually an oriented incidence matrix. > > >> Thus if you indeed change this behaviour, it may make sense to add a >> method that imposes an arbitrary ordering. > > For discussion purposes, let me toss out some ideas on improving the > interface for computing incidence matrices. For a graph object G, I > propose we add a keyword to G.incidence_matrix() so that the interface > is now changed to the following method signature: > > def G.incidence_matrix(orientation=False) > > The keyword "orientation" takes on a Boolean value. So > G.incidence_matrix(orientation=False) or G.incidence_matrix() returns > the unoriented incidence matrix of an undirected graph G. Furthermore, > G.incidence_matrix(orientation=True) returns the oriented incidence > matrix of an undirected graph G, which is the current behaviour. The > keyword "orientation" has no effect if G is a digraph. So > "orientation" is only meant to affect undirected graphs. > > Let's consider whether or not to leave the method incidence_matrix() > in the module generic_graph.py. I consider it more appropriate for > graph.py (a module for undirected graphs) to deal with incidence > matrices for undirected graphs. Similarly, digraph.py can deal with > incidence matrices for digraphs. At the moment, incidence_matrix() > resides in generic_graph.py, a module with over 10 thousand lines. > > I would like to open a ticket to implement the above proposal or any > good proposal that comes up with regard to improving > incidence_matrix(). > > Thoughts?
This seems reasonable to me. I will act as a referee but would prefer to have at least one other (Rob Beezer?). > > -- > Regards > Minh Van Nguyen > > -- > To post to this group, send an email to sage-devel@googlegroups.com > To unsubscribe from this group, send an email to > sage-devel+unsubscr...@googlegroups.com > For more options, visit this group at > http://groups.google.com/group/sage-devel > URL: http://www.sagemath.org > -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
