R help - Jun 2007 - Boxplot issues

Boxplot and bxp seem to have changed behaviour a bit of late (R 2.4.1). Or maybe
I am mis-remembering.

An annoying feature is that while at=3:6 will work, there is no way of
overriding the default xlim of 0.5 to n+0.5. That prevents plotting boxes on,
for example, interval scales - a useful thing to do at times. I really can see
no good reason for bxp to hard-core the xlim=c(0.5, n+0.5) in the function body;
it should be a parameter default conditional on horizontal=, not hard coded.

Also, boxplot does not drop empty groups. I'm sure it used to. I know it is
good to be able to see where a factor level is unpopulated, but its a nuisance
with fractional factorials and some nested or survey problems when many are
known to be missing and are of no interest. Irrespective of whether my memory is
correct, the option would be useful. How hard can it be to add a
'drop.empty=F' default to boxplot to allow it to switch?

Obviously, these are things I can fix locally. But who 'owns' boxplot so
I can provide suggested code to them for later releases?

Steve Ellison



*******************************************************************
This email and any attachments are confidential. Any use, co...{{dropped}}

>>>>> "SE" == S Ellison <S.Ellison at lgc.co.uk>
>>>>>     on Fri, 22 Jun 2007 13:02:20 +0100 writes:
    SE> Boxplot and bxp seem to have changed behaviour a bit of late (R
2.4.1). Or maybe I am mis-remembering.
    SE> An annoying feature is that while at=3:6 will work, there is no way
of overriding the default xlim of 0.5 to n+0.5. That prevents plotting boxes on,
for example, interval scales - a useful thing to do at times. I really can see
no good reason for bxp to hard-core the xlim=c(0.5, n+0.5) in the function body;
it should be a parameter default conditional on horizontal=, not hard coded.

    SE> Also, boxplot does not drop empty groups. I'm sure it used to. I
know it is good to be able to see where a factor level is unpopulated, but its a
nuisance with fractional factorials and some nested or survey problems when many
are known to be missing and are of no interest. Irrespective of whether my
memory is correct, the option would be useful. How hard can it be to add a
'drop.empty=F' default to boxplot to allow it to switch?

    SE> Obviously, these are things I can fix locally. But who 'owns'
boxplot so I can provide suggested code to them for later releases?


Legally speaking, I think that's a hard question the answer of
which may even depend on the country where it is answered.
I would like to say it is owned by the R Foundation.

Suggested improvements of the R "base code" should be made and
discussed on the R-devel mailing list. That's exactly the main
purpose of that list.  
Such propositions typically make it into the code base
if you are convincing and you provide code improvements that
convince at least one member of R core that it's worth his time
to implement, document, *and* test the changes.

Also, as on R-help, it helps to work with small reproducible
(ideally "cut-n-pastable") R code examples.

Regards,
Martin Maechler

    SE> Steve Ellison