I’m using Region3s to detect parts. The reason why it didn’t paint the walls on the second bomb was because I didn’t mark those walls as paintable.
And yeah, I’m using part reductions from the previous post. That blast has a radius of roughly 12 studs, and the grid is 0.25 studs in size. That means I’m painting roughly pi*(12^2)/(0.25^2) = 7238 cells per explosion.
The color of each cell is stored in a flat array. The painting of each cell is done by the server. To reduce the amount of data I send to the client and to reduce the amount of work the client has to do, I apply the algorithm that sircfenner described in this post to convert the cells into a bunch of rectangles. I encode each rectangle as three characters (a color index, the top-left corner of the rectangle, and the bottom-right corner of the rectangle) which I broadcast to each client.
(I’m not sure what minimal rectilinear polygon rectangulation is, but from the googling I’ve done it doesn’t seem like I’m doing anything of the sort.)
So um, a little bit ago I ended development on Space Frontier, and have been working on a new project.
Don’t wanna give out too many details on the project yet, but if you guys know anything about Space Frontier I’m sure that you know it has some pretty terrible GUIs. Been trying to work on that for this next game.
Not that it is super technically impressive or anything, but here’s my hp bar, hunger bar, and timer guis for the game so far.
I may have a suggestion to help with performance on lower tier machines while allowing higher tier machines continue with the detailed effects.
Checking base on their graphic level settings, you can either generate detailed parts as the paint or generate a fewer parts with a BillboardGui to imitate the detailed effects as a visual illusion.
Just a suggestion, I want to know what do you think.