seq --> Dealing with bugs in R, etc.

Prof Brian Ripley Prof Brian Ripley <>
Tue, 15 Sep 1998 17:53:28 +0100 (BST)

> Date: Tue, 15 Sep 1998 09:44:15 +0200
> From: Martin Maechler <>
> To:
> CC:
> Subject: Re: seq  --> Dealing with bugs in R, etc.
> >>>>> "Jim" == Jim Lindsey <> writes
>    in a private mail to me
>     Jim> That bug in seq() really has shaken my faith in R. How
>     Jim> many other known bugs are there that we on r-help are not informed
>     Jim> of and which can falsify our statistical analyses??? It is
>     Jim> imperative that patches for such things be made immediately
>     Jim> available on r-help. Jim
How does that `falsify a statistical analysis'? It just gives a vector
that is slightly too long. I cannot imagine this is the worst bug
in R right now.

> I think the answer to this should be known/discussed more widely:
> Jim, you are right: 
> The patch to seq(.) should have made it to  R-help.
> [[actually in the mean time, the issue was raised that my patch
>   was much more than was needed for dealing with rounding errors
>   in the  seq(.,., by= ) case
> ]].
> Note that there *are* other known bugs, all(?) of which being 
> of much more low-level (and ``deeper'') nature.

Not all: e.g. mahalanobis() gives incorrect results almost all the 
time in 0.62.3, and glm with offset is wrong often too (and both
those are in the R-bugs database).

And seq gave wrong answers in the previous releases too: 
seq(1.575, 5.125, by=0.05) was wrong. At least the simpler uses are
now correct. 

I do think R-bugs is the right place for bugs.

Brian D. Ripley,        
Professor of Applied Statistics,
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272860 (secr)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

r-devel mailing list -- Read
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: