Hi, Am Dienstag, den 10.02.2009, 11:59 +0100 schrieb Heinrich Apfelmus:
However, I would be surprised if someAlgorithm could not be formulated directly on the tree or at least satisfies a few invariants like for example
map fst . someAlgorithm = map snd
Also, how does cutTreeApart arrange the list? The idea is that most of the tree structure survives in the list and can be reconstructed.
Probably not. My algorithm calculates the amount of light that shines through the branches, and the amount of light cought by the branches (seeing branches as approximations for leaves here :-)). The algorithm works by taking the list of branches (represented by their start and end point), projects them along the direction of light, creates a list of start and endpoints, sorts them (to be able to sweep the line somewhat efficiently) and adds the projections of all intersections of branches. Then it iterates through the intervals on this line, gets the list of branches that are hit by this interval, sorts them by hights and adds, from top to bottom, the appropriate, decreasing portion of the light that comes in this interval. In this process, the branches are sorted around quite a bit, and I assume it would be hard to preserve the structure. If you want to see code (not sure though if you really want to see that code :-)), it’s in http://git.nomeata.de/?p=L-seed.git;a=blob;f=src/Lseed/Geometry.hs Lines 74-142. Greetings, Joachim -- Joachim Breitner e-Mail: mail@joachim-breitner.de Homepage: http://www.joachim-breitner.de ICQ#: 74513189 Jabber-ID: nomeata@joachim-breitner.de