Forbes Magazine, Names You Need to Know in 2011: R Data Analysis Software: http://blogs.forbes.com/smcnally/2010/11/10/names-you-need-to-know-in-2011-r-data-analysis-software (and links therein)
And some of its developers are suggesting that they scrap it and start over (don't know if the whole "Core Team"'s on board tho'): http://www.stat.auckland.ac.nz/~ihaka/downloads/Compstat-2008-Slides.pdf
Are there parallels like this in the development of other languages/environments/ecosystems (e.g., Python3, Perl6 "revisions")? How do these efforts usually end up (I guess we're still waiting to see about Python3 and Perl6...) -- and how would it affect your business's decision to develop a library in this environment?