Proposed by Mathieu Dehouck (profile, biography)
How will I do that project
I'll spend at least 25 hours a week on the project and see with my mentor if the work I'm doing is what was expected and is useful.
I will send mail on moose dev each time I have something interesting to show.
What methodologies will I use
I will have a look at the solutions used by other visualization tools, and see how them can be used with Roassal.
I will also look into Roassal to see if there are simple optimizations to save time, such as just changing "translateto: " into "position: ", it save up to one seconde (from 5638 to 4656) on a 130 elements graph, so on bigger I let you imagine.
I will do interesting commits to show my work to the comunity, and I will use "timeToRun" and tools such as it very much.
Suggested timeline and milestones
The first thing to do is to have a good quad tree so that we save time when moving and displaying the elements.
Then will seek for optimizations.
Where I see the risks
Even if we can save time by making optimizations here and there, the goal is also to provide "generic" solutions to globaly improve Roassal scalability, and not only on a layout or the other.
The main risk is to have to high expectation, because even with good improvements, we will always have to wait the computer with big complex graphs.
How the results will look like
Roassal, but faster.
|