R help - Mar 2010 - mixed effects models for grouped time survival data

How do I fit a mixed effects model with two crossed random effects for grouped
time survival data?

I tried coxme with no luck.

Suppose that y is survival time, uncens is censoring indicator, trt is treatment
below.

ran.eff1 and ran.eff2 below are two crossed random effects. This way, it 
wouldn't run. If I remove one of the random effects, it runs perfectly.

coxme(Surv(y,uncens) ~ trt + (1|ran.eff1)+(1|ran.eff2),data=mydataframe)

	[[alternative HTML version deleted]]