Nobody doubts that R (r-project.org) is a great statistical language and analysis system, but hey, developers, could you please …
1) create a modern website (the present one is so old school that I wonder why you did not use animated gifs for bullet points) … just look at what the bioconductor guys do at bioconductor.org
2) set read.table(…, stringsAsFactors=FALSE) as the default
3) set sum(c(NA, NA), na.rm=TRUE) to NA (not 0), because if I don’t know how much money I have and I don’t know how much money you have, than I cannot conclude that we both have no money (see!)
4) don’t round summary() output because this is highly misleading
summary(c(1:12345))
Min. 1st Qu. Median Mean 3rd Qu. Max.
1 3087 6173 6173 9259 12340
5) update the ugly R graphical user interface (design is function!)
6) include a decent text editor as a default (syntax highlighting and stuff)
7) provide a simple, cross-platform batch command which integrates the script, output, and errors in one file (and allows to forward user defined variables to the program)
8) avoid using the annoying TCL/TK windows when installing packages from the console (who needs a graphical user interface anyway)
9) separate the process of the console and the editor (if R crashes you don’t want to loose your last edits of your script)
10) provide a Latex style file for package reference manuals which automatically wraps long lines of code. Obviously, the current Latex style files cannot wrap long lines of code examples and just crops them at the end of the line.