[Rd] improved pairs.formula?

[Berwin A Turlach]

Dear Brian,

>>>>> "BDR" == Prof Brian Ripley <ripley at stats.ox.ac.uk> writes:

    [...]
    BDR> Still not a description, just two examples.
O.k., I believe that I slowly understand want you want.  So in my book
a description of "the job" would be: "that it is allowed to use '.' and
'-' in the formula given to pairs thus allowing pairs.formula".

    BDR> Since later on you took a description to be an example, I see
    BDR> the confusion.
Just to avoid future confusion, so how does one distinguish a
description from an example, especially if the description is after
the phrase "such as"

    BDR> and why you chose such an unusual piece of code to do it?
    >> Mmh, I don't think of it as being so unusual, most of it was
    >> gleaned from other R function.  Well, I realise that R
    >> programing paradigms change over the years, so I must have
    >> gotten them from quite old routine.
    BDR> I guess you got it from an S not R function.
Quite possible.  Sorry, I can't check since I never had access to S.

    BDR> (E.g. what is the prescription for the ordering of terms, and
    BDR> why do you think the rownames of the factors and the
    BDR> variables might be in different orders?  They are set the
    BDR> same in the C code.)
    >> In case that a user foolishly specifies a more complicated
    >> formula having not read the help pages.  It seemed to me that
    >> this was the only construct to figure out which variables are
    >> actually appearing in terms of the formula.
    BDR> Really?  Please check what I wrote: `variables' and `rownames
    BDR> of the factors' are always the same, apart from the response.
    BDR> Please show an example where you got something different.
Sorry, seems to be some misunderstanding here.  With 'variables' you
mean the "variables" attribute of the object extracted by "mt <-
attr(mf, "terms")" in my code?  It is well possible that this is
always in the same order as the rownames of the factors but I didn't
find that in the documentation and I didn't study the C code to write
this function.  But I take your word for it and in this case I can see
how the code could be simplified.  Shall I submit a revised version or
not bother?

    BDR> There are several places where only the allowed form of
    BDR> formulae is specified in this way.
In this case, may I request that this places are better marked than
they are at the moment? 

    BDR> You are not allowed interactions, for example, and it refers
    BDR> to `each term'.
Indeed, it states "Each term will give a separate variable in the
pairs plot, so terms should be numeric vectors".  So some users might
be surprise that the following works:

> dat <- data.frame(x=rnorm(30), y=rnorm(30), f=rep(c("A","B", "C"), 10))
> pairs(~x+y+f, dat)

Who, but the most arduous student of R help page language would have
thought that dat$f is a numeric vector?  :-))

    BDR> `.'  is not documented to work (and used not to).
So the question is whether it should remain this way or not.  But that
is for the R core team to decide and I shall stop my contributions to
this thread now.

Cheers,

        Berwin

========================== Full address ============================
Berwin A Turlach                      Tel.: +61 (8) 6488 3338 (secr)   
School of Mathematics and Statistics        +61 (8) 6488 3383 (self)      
The University of Western Australia   FAX : +61 (8) 6488 1028
35 Stirling Highway                   
Crawley WA 6009                e-mail: berwin at maths.uwa.edu.au
Australia                        http://www.maths.uwa.edu.au/~berwin