# demonstrates the need for a semantic layer for the data model.  Kludgy code in 
# AP requires special knowledge to get this to work.  X,Y,Z should be a dependent 
# parameter on implicit parameter T.

x= randn( 10000 )
y= randn( 10000 )
z= sqrt( x**2 + y**2 )
y.putProperty( QDataSet.DEPEND_0, x )
y.putProperty( QDataSet.PLANE_0, z )
result= y

x= randn( 10000 )
y= randn( 10000 )
z= sqrt( x**2 + y**2 )
result2= link( x, y, z )  # this is the new way of doing this, but there's a bug so it's not handled properly.