[R] specify the number of decimal numbers

Wacek Kusnierczyk Waclaw.Marcin.Kusnierczyk at idi.ntnu.no
Thu May 14 17:57:30 CEST 2009

James W. MacDonald wrote:
> Wacek Kusnierczyk wrote:
>> do they make pompous claims about their software and disregarding claims
>> about others' as well?
> My mistake. I thought your concern was for the quality of the software
> (quality of course being defined by a certain committee of one). But
> it appears this is of a more personal nature, so I'll let you get back
> to tilting at windmills.

your mistake.  the point was, claims have been made here to the effect
that (a) for rounding off a 5, the IEC 60559 standard is expected to be
used, (b) round in r complies with the standard, (c) it is rare for
round() to be called explicitly in R code: rounding is usually going on
inside print routines.  in addition, (d) not complying with the standard
was called 'excel bug'.

as you can clearly see from the examples, something is wrong with (a) +
(b) + (c), and that's what i raise as a non-negligible issue.

i haven't seen perl folks making any claims of the (a)-(c) sort, and
neither being rude the (d) style.  (it does not mean, of course, that
they haven't made such claims.)  i have no idea why i would have to
complain to perl people about an apparent logical inconsistency in what
the chief r core developer says;  do you?

i also have no idea why you would need to try to turn my well-grounded
remark into tilting at windmills;  if that's not silly, something's
clearly wrong with me.


More information about the R-help mailing list