[PD] new pmpd developer

[cyrille]

hello Johannes,


Johannes Taelman wrote:
> cyrille wrote:
> 
>> so, if anybody have idea regarding pmpd, it's time to share it...
> 
> 
> Hi Cyrille,
> 
> First of all I've only briefly used pmpd, and didn't have the time to 
> delve into the depths of it. So my comments may be off.
no problem, any coment are welcom.

> But for wat it's worth:
> 
> 1. Non-spherical objects and object rotations. More or less like 
> OpenDE's possibilities: 
> http://opende.sourceforge.net/ode-latest-userguide.html

well, in fact pmpd use only ponctual mass : this mean that all masses 
does not have any volume. they are not spherical.
there reprensentation can be whatever you wish.

intruducing rotation would mean introducing shape etc.
the equations then became to much complex to solve them in an easy way.
so, using ODE or other 3D dynamic software is much more complex than 
using pmpd.

you can also use breve3D ( http://www.spiderland.org/breve/index.php ) 
if you wish an openGL + physical modeling tools. bridges with pd can 
certainly be made.

doing such simulation tools is a very huge work, i don't think i'll be 
able to do it someday.

> 2. Break the one-pd-object to one-physical-object relationship. I often 
> iterate GEM geos using [repeat], to avoid 100's of instances of an 
> abstraction.  I don't have a clear vision of how this could be
> implemented, I understand that there is a need to allocate the objects 
> somehow in order to simulate them. Maybe pmpd could offer an object that 
> allocates arrays of physical bodies and links.
this is exactly what the [pmpd] object is doing ( part of pmpd v0.06).
Nicolas will work only on this aspect of pmpd.

> 3. Allow multiple environments. Now the namespace is global, loading 2 
> pmpd example patches clashes.
clashes does only occur in the send / receive. This can easilly be solve 
  using $0-name. (it a problem with my exemple, not with pmpd).

> 4. Tighter GEM integration. Maybe the GEM transformation (translateXYZ, 
> rotateXYZ) handling could be fitted into a pmpd-object?

why don't you like the curent solution of using abstraction?

i personally use gem, but pmpd should also be usable for peoples using 
data structure, pdp, gridflow, framstein, xgui, or any other 
visualisation tools.
pmpd only compute movement, that you can use as you wish.


> Preferably also 
> compatible with gem-repeat constructions.
it can not be so easy.
i don't see any solution for the pmpd objects to be compatible with the 
gem-repeat construction, but thanks to pd flexibility, you can get GEM 
to be compatible with [pmpd] object.

ok, wait for the exemples....

thanks
Cyrille



> 
> Johannes
> 
>