[R] What made us so popular Nov 16-20?
ggrothendieck at gmail.com
Tue Nov 29 03:04:07 CET 2005
On 29 Nov 2005 01:45:34 +0100, Peter Dalgaard <p.dalgaard at biostat.ku.dk> wrote:
> Seeliger.Curt at epamail.epa.gov writes:
> > Duncan asks:
> > > Did we get mentioned somewhere (e.g. Slashdot), or was someone just
> > > experimenting with some automated downloading?
> > R was mentioned in last week's (I think) O'Reilly newsletter, which
> > included a link to a short article showing how easy it is to get R to
> > graph stuff like stock price histories. That's the publisher, not the
> > talking head.
> [See other mail for the link]
> > For what it's worth, the article isn't worth chasing down. It left a
> > beginner like me disappointed that R's capabilities weren't better
> > shown, and that he relied on Perl to do data manipulation.
> Er, where did you see Perl being used? The only thing that irked me
> (admittedly, I only skimmed the article) was that he was using
> regression models to test for correlation (why not cor.test()?) and
> speculates a bit wildly about the sign of a clearly nonsignificant
> relation. It's a bit superficial, but I suspect that this sort of
> paper has to be.
That article not only gave the (false) impression that one
needed something like perl to do data analysis but it
failed to use time series objects to represent time series
making the examples poor (e.g. x axis was labelled
1, 2, ... instead of using time) and unnecessarily lengthy
By the way, during November there was also a vote on
alt.comp.freeware for the pricelessware collection
(which R failed to get on in 2005 but did manage to make it on
for 2006) and this may have been the reason, in part, for
More information about the R-help