Re: Nested interval tree encoding

On Jul 3, 2:37 pm, "Eric DeCosta" <edeco..._at_mathworks.com> wrote:
> In my particular exploration of this topic, I do, indeed have all four
> elements of a 2x2 matrix stored

Then, applying the extended Euclidean algorithm is unambiguous, right? The sigmod paper p.4 illustrates how to expand the matrix by adding more columns. The algorithm stops as soon as we add column [0,1].

> -- and the right column is the continued
> fraction of the node's parent; so I don't think the properties of adjacency
> list are particular to Farey fractions.

I'm not sure I understand your comment. Sure the left column [3,4] of the child
[3 5]
[4 7]
refers to the right column of the parent [2 3]
[3 4]
In this sense it is similar to adjacency list. The difference, of course, is that in matrix encoding we have a very strict policy what values are admitted, and in adjacency list you can label tree nodes by anything. Received on Fri Jul 04 2008 - 00:31:02 CEST