TaaRtJoin Date: 2009-04-26
Post Count: 5070 |
Hmm, well I feel the essence is in the physicless instance. Whether you make a polygon out of really doesn't need to be an Instance of its own in my opinion since merely changing the size and shape shouldn't be a problem to begin with (as there's hardly and expense to it) |
|
|
There will be less script strain from calculating the positions and orientation of the parts relative to the nodes in the case of the triangular plane instance compared to the physics-less part. A good studio update would bring both in any case.
#code print('Goodbye world') |
|
TaaRtJoin Date: 2009-04-26
Post Count: 5070 |
Attachments themselves require a physical part though, where a physicless Instance can be entirely on its own. It's not a lot of strain on a script if a position and orientation is set of a physicless part to begin with (and it won't be multiple nodes, just a single CFrame) |
|
|
Well that's where the physics-less part is different. As I said in my op, I'm thinking of something more along the lines of a trail or beam instance, just changed to make a triangle between 3 attachments. They can't really compare with a physics-less part.
#code print('Goodbye world') |
|
|
I can house my 25 x 25 array of attachments in one single part and have them move independently (as I already have)
It's just the polygons to connect to these attachments and make a wavy surface, using the array as nodes.
Physics-less parts can't do this as efficiently is what I was saying.
#code print('Goodbye world') |
|
soutenuJoin Date: 2011-12-09
Post Count: 1021 |
it'd be easier to just add a property to basepart which would decide whether physics are calculated or not |
|
TaaRtJoin Date: 2009-04-26
Post Count: 5070 |
Adding and removing a part dynamically from the event connections and such would be (much) more problematic than a new Instance that always lacks these connections to begin with, soutenu |
|
|
neat idea but cancollide false anchored parts can be used to have a similar effect, the most I have tested with was 1 and a half million parts and after they finished loading in there was no lag. |
|
TaaRtJoin Date: 2009-04-26
Post Count: 5070 |
@128 those are still connected to a bunch of events, if you'd actively start moving them you'll notice a difference to when they'd not have these connections |
|
|
Well they can't move
Because they are anchored
Unless you mean with a script, in which case I'd say 1 and a half million anything being changed will cause lag. |
|
TaaRtJoin Date: 2009-04-26
Post Count: 5070 |
The idea is that 1 and a half million Instances without connections to events and all the good stuff is less laggy to set the CFrame for than the same idea with existing (anchored) parts |
|
|
For sure, was just saying that unanchored cancollide false parts are the closest alternative currently. |
|
|
I think this exact suggestion would be too specific. If we just had a part that wasn't affected by physics and all that jazz, many more people could find it practical. For your specific request you could just make a sort of plugin with the new instance. |
|
|