R help - Aug 2004 - summary output for inverse Gaussian GLM

I'm getting inconsistent output about the link function from the
summary() command when fitting an inverse Gaussian GLM:


 > summary(glm(ig.formula, family=inverse.gaussian(link = "log"),
+ data=mydata, start=start.vals))$call
glm(formula = ig.formula, family = inverse.gaussian(link = "log"),
     data = mydata, start = start.vals)
 > summary(glm(ig.formula, family=inverse.gaussian(link = "log"),
+ data=mydata, start=start.vals))$family

Family: inverse.gaussian
Link function: 1/mu^2


Has anyone else run into this problem?
I'm running v1.9.1 in Windows.

    E.J.

On Tue, 10 Aug 2004, Edward Dick wrote:
> I'm getting inconsistent output about the link function from the
> summary() command when fitting an inverse Gaussian GLM:
>
Yes, it's a bug in the name, not in the result, fortunately.  The
inverse.gaussian family object always has the name of the link set to
1/m^2, regardless of the actual link function.

	-thomas