Comments by Mesrop Simonian

Mesrop Simonian 24-Jun-11 13:09pm View

Now that definitely looks like it would work. I'm going to try it now.

Mesrop Simonian 24-Jun-11 13:08pm View

the SetNewTotal method always returns 0s... also, isn't total at some point supposed to be assigned the value of r?

Mesrop Simonian 23-Jun-11 19:36pm View

Absolutely. You may be on to something here, because I think you have a partial solution for POP, but POT is still a different issue. I compared the edited solution to the previous version and I only saw differences in the IncrementChildren section. So this IncrementChildren method, I assume, is on branch level, not tree level. If so, where are POP and POT? Branch level or tree level? POP can be branch level and this solution might work, but POT cannot, because "root" is not known at branch level (out of scope). And if that's the case the solution is not implementable. Do you see why?

Mesrop Simonian 23-Jun-11 18:49pm View

Bottom up vs top-down doesn't matter I think. The question is the ability to do horizontal navigation as well - we got the up/down we need the left/right, so instead of a depth-first search, it needs to be a breadth-first search, but in that case I don't know how to ensure that every single branch in the structure will be visited.

Mesrop Simonian 23-Jun-11 18:42pm View

MainBranch I meant to say

Mesrop Simonian 23-Jun-11 18:41pm View

Main Branch I meant to say

Mesrop Simonian 23-Jun-11 18:41pm View

I'm calling it in the Tree object's MainRoot property's set method, so it's supposed to scan all the roots contained within MainRoot and assign the POT and POP properties for each root.

Mesrop Simonian 23-Jun-11 18:15pm View

single root node that is the parent of everything.

Mesrop Simonian 23-Jun-11 18:14pm View

This solution unfortunately won't work because root.TotalOfChildren cannot be evaluated without recursive calls. When a branch is added, total number of children is incremented only by 1. However, the added branch may have a number of children of its own (say 10), which should be accounted for in the POT evaluation. In this case, the calculated POT for the main branch that contains the entire tree will be 1 (me) + 1 (my children) over 12 (total number of branches) = 16%, which is incorrect if all 10 children are part of the parent's child. The right solution must come out to 12/12 * 100 = 100%. I really don't think this problem will be solvable without recursive calls to the evaluation methods both for POT as well as POP.

Mesrop Simonian 23-Jun-11 15:55pm View

thanks, but my tree is built inside the code itself, although subsequently it will be in the database... but right now, i need it to work without a backend.

Mesrop Simonian 23-Jun-11 15:54pm View

% of total equals The current branch + all of its children / total * 100, % of Parent is same with regards to the parent only instead of the big total.

Mesrop Simonian 23-Jun-11 15:46pm View

Thanks for the answer, but I completely screwed up the question. Maybe you'll be up for this challenge as well - the problem isn't that children of children are not being processed, they are. The problem is coming back up in the hierarchy and navigating to siblings, so that ALL branches end up getting processed through the method. Because right now it goes down the list as long as children have children, but then it stops, so if the main branch had 2 subbranches, only the children of the first sub branch are visited, then the execution ends. Sorry for the confusion, I myself wasn't sure exactly why the calculations were wrong.

Comments by Mesrop Simonian (Top 12 by date)