On 27/07/12 20:55, David Winsemius wrote:
On Jul 26, 2012, at 10:06 AM, FJ M wrote:
It would be a useful additon to the help page to add
integrate(dnorm, lower = -1.96, upper = 1.96, mean = 2, sd = 1)
Wouldn't most statisticians instead use the more accurate and
undoubtedly faster:
pnorm(1.96, mean=2, sd=1) - pnorm(-1.96, mean=2, sd=1)
... or p_xxx where _xxx is distribution of interest?
I conjecture that the OP was just using "dnorm" as an easily
reproducible example,
and was really interested in a function or functions for which no
built-in "indefinite
integral" exists.
(And I do not see that every help page needs a worked example of the
proper use of the 'dots' argument. That is a basic R lesson and it is
only by chance that you have needed to learn it in the context of
integrate(). A large fraction of other functions offer that facility.
You might review the material you used when learning R to see at what
point you skimmed over that crucial topic too quickly. )
Well, yes, but ................. the use of the "..." argument to
pass arguments to another
function (which itself is an argument) is a bit subtle. At least
initially, to those of us
who are Bears of Very Little Brain. A wee example would be
redundant, as your discussion
implies, but a little redundancy rarely hurts and tends to add
robustness.
We can't expect the documentation to explain "..." every time it
appears as a function
argument, but giving an example in "important" instances (with
important being defined
as when someone requests an example) seems to me to be a good idea.
cheers,
Rolf
P. S. My Thunderbird mailer does not recognise the word "Winsemius" as
a word and
suggests "inseminates" as a possible alternative. :-)
R.
P^2. S. Strangely enough, Thunderbird does not recognise
***Thunderbird*** as a word!!!
R.
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
