R help - Jan 2010 - piecewise regression using nlme function

Is a 95% CI on a breakpoint fixed effect legitimate when the nonlinear
equation is continuous but not differentiable at the breakpoint?
 
I used the nlme to generate a non-linear mixed-effects piecewise model
with initial slope zero prior to an unknown breakpoint. It did give a
95% CI for the breakpoint.
 
Thanks
Erik
 
whole model is complex, but here is the simple equation used:
 
dr.hockey<-function(a,b,x,brk){
 a+ifelse(x<=brk,0,b*(x-brk))}
 
 

	[[alternative HTML version deleted]]