Yeah it's not that bad; it's comp-cost equation. The algorithm is truly recursive. So every IF or SUM that we can removed is removed time the number of items in the outline. That's why I don't like adding orthogonal feature. Each option that has to be test globally mean and extra branch in the algorithm has to be run through for every node in the task tree.
So you have to think smart. The current rev I was able to identify 5 items we use to calc on the fly because it made sense. I was able to move those over it the Task OBJECT and handle the updating process in the gui. That removed something like 12-20 operations from the loop and that is per Node iteration; that's a ton; We also probably remove 15-20% of the main IF/Then branch with the new matrix and case statement. Part of this is that the algorithm got "extend" in version 1.5; and should have been re-factored at REV2, there wasn't time to it go bloated. It's now lean and mean and able to grow again; BUT we need to try a running version of REV3 and have people see how it behaves as designed before doing anything else. This is all based on the desktop code of course; I've never seen the source code for the PPC version; the team takes the desktop and rewrites it for the PPC so they might have already made many of the same improvements in the past; they weren't anything special they just require groking the algorithm and having the time to dedicate to it. On Jul 16, 4:06 pm, "Vallon, Justin" <[email protected]> wrote: > Bob, > > You might have considered this, but if you are dynamically computing the > scoring, you could try a cached-value approach, where you cache the > calculation of the score, and some operations (adjusting inputs, > reparenting, modifying weighting of parent) would invalidate the > calculation. If you are lazy about recalculating the score, then it > would be no slower than now on display (with a speedup after being > computed once), for some additional cost when you have to invalidate the > tree (when the root is modified or reparented). > > Of course, this is a space-for-time tradeoff, and the devices are > memory-constrained. > > -Justin > > > > -----Original Message----- > From: [email protected] > > [mailto:[email protected]] On Behalf Of scoobie > Sent: Thursday, July 16, 2009 2:42 PM > To: MyLifeOrganized > Subject: [MLO] Re: Computed-Score Priority and Start Dates > > Bob, > What's your take on being able to do all this in an iphone sized > processor and screen? > Do you think its possible? --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "MyLifeOrganized" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/myLifeOrganized?hl=en -~----------~----~----~----~------~----~------~--~---
