Does instantiating A LOT of Max Objects from JS or JAVA makes sense here …??
those who have followed my audio & visuals system incursion probably already thinking I’m insane.
totally true. on an art & technology point of view, of course :)
Looking for the best way to:
- design visuals little objects instantiated multiple times in a 3D space
- keep track of each object in a central core (JS, JAVA , whatever) in order to send them messages, to save/restore the whole of them
I first went on the road of JAVA (or JS) instantiating directly inside of them all the objects in my patch (using JitterObject instantiation etc).
It seems very expensive for the CPU. I didn’t dig it especially during months but intuitively, I feel it isn’t the good way and it can be very awful to take a jitter patch, reproduce it totally using obj.maxtricalc(matrix,matrix) things like that…
So, I thought about something like a poly~ for jitter.
Read a lot about some implementation etc. But I’m not convinced.
indeed, if in a map (=a preset of my 3D world) I would have 3 or 4 huge grid to display, I could have around 50 other type of objects and poly~ would be killed probably.
So… I thought about another way.
Create one abstraction per object would really be a nice way. I’m working like that every time I need to instantiate patches I’m reusing.
Considering I could easily, using that way, add some specific custom properties in the abstractions, track each abstractions specifically (name)
I could create a huge JAVA core (for fast stuff) that could instantiate the abstractions directly.
It would be an insane patch: at loading time, the JAVA would create ("physically") object in the patch, placing them everywhere in some subpatches.
I could: communicate with each one of them via JAVA, listen to them from JAVA (I guess by linking them programmatically to a huge hub going into the JAVA inlets)
I could also save all data to XML from the JAVA core, restore data from XML too.
And especially, I could take benefits of graphical programming and all Max/Jitter objects speed… which is just basic & obviously GREAT.
- does it make sense ?
- would that way efficient compared to stuff instantiated/encapsulated inside JAVA ?
- would you see any cons or limitation around this ?
This sketching test travel is amazing and hardcore these days.
Many thanks in advance :)
just to add something more:
I played a bit with JS instantiating abstractions.
it works very fine, obviously.
it could be the way to follow for me.
In each abstraction, I would have the patch of a type of object to display.
still digging and VERY interested by your opinions.
I’d be very interested by Cycling74′s point of view about especially PROS / CONS of duplicating A LOT of abstractions optimized and containing only Jitter objects (I mean no JAVA or other externals) versus instantiating all of that inside JAVA.
I’m refining my system and it appears there is A LOT of PROS for that way.