The following code snippet taken from plot.survfit now gives me a surprising
result:
tempx <- c(0, 400, 0, 0)
ytemp <- c(.0657, .98, 1, 0)
plot(range(xtemp), range(ytemp), log='y', type='n')
In that the max value for the x axis is now 10. The 0 on the end of ytemp was
actually a
mistake and shouldn't be there, if you remove it then the upper limit is
sensible.
plot(range(xtemp), range(ytemp[1:3]), log='y', type='n')
I've now fixed this in the survival package. I can't say when the odd
behavior crept in,
but am sure it was not always there. I draw log scale plots only a few times a
year.
Terry T.
