[R] Debugging R's code: boxplot.stats

[Martin Maechler]

>>>>> "Duncan" == Duncan Murdoch <murdoch at stats.uwo.ca>
>>>>>     on Sat, 28 Oct 2006 08:03:33 -0400 writes:

[...........]

    Duncan> Not sure why boxplot.stats is in grDevices rather
    Duncan> than graphics, but that's not really relevant to
    Duncan> your problems.

because it does "computation for Graphics" (capitalized on purpose)
which should not at all be limited to old-style S graphics which
is the topic of the "graphics" package.
E.g. "grid" and grid-based graphics should use boxplot.stats
(for consistency).
For the same reason, "color computations" are in 'grDevices',
[[and we have acknowledged a while ago that 'grDevices' has become a
  slight misnomer

Martin Maechler