Chris Hill wrote: > Shouldn't we have a policy of giving a nice error message > for parameters that are retired/moved? > The parameter has been in "data" since checkpoint35, so there > will be a few people using it. Yes - I agree we should. A.