[Rd] argument handling in boot.ci

[Duncan Murdoch]

On 23/03/2011 11:35 AM, Matthew Vernon wrote:
> Hi,
>
> boot.ci fails to check if the "type" argument is valid or not. This
> tripped me up when I had a script that called boot.ci(type="normal",...)
> repeatedly - only later when I tried to do some further analysis did I
> realise I'd got nothing useful back from boot.ci.
>
> It strikes me that this behaviour is probably incorrect, and that
> boot.ci should stop if the "type" argument (or, indeed, any other
> supplied argument) is invalid. Am I correct?


boot.ci is in a contributed package (it's "Recommended", which means 
that a copy is distributed with R, but it is not part of R), so 
suggestions should go to the maintainer.  But you need to put together a 
simple example to illustrate the problem.  When I modify the first 
example in example(boot.ci) to specify type="normal", I get an error 
message.

Duncan Murdoch