[Verse-dev] Verse specification: transformation matrix
Jiří Hnídek
jiri.hnidek at vslib.cz
Sat Jul 23 11:14:28 CEST 2005
Hi,
I try to implement sending and receiving object node transformation
matrix. Current specification of verse protocol isn't too much useful.
Why? I will never need speed and acceleration of position/speed, but
1) I have to do crazy converting from transform matrix to pos[3] ,
rot[4] and scale[3] vectors ... send each vector separately.
2) When I receive any change of object node transformation, then I have
to do crazy converting from pos[3], rot[4], scale[3] to transformation
matrix (several matrix multiplication)
3) I can't transfer information about shear, etc.
4) I guess, that other clients (Maya, Max) will have to do same. They
will have to do similar unnecessary operation (in plugins or clients
will do it internaly)
Conclusion: verse specification offeres many unnecessary parameters,
complicate implementation, miss some functionality.
My proposal: add these commands:
void cb_o_transform_matrix_real32(VNodeID node_id, real32 *mat);
void cb_o_transform_matrix_real64(VNodeID node_id, real64 *mat);
These commands would be used for sending transformation matrix:
real32 mat[4][4];
real64 mat[4][4];
Best regards,
Jiri
More information about the Verse-dev
mailing list