How do you program statistics?
In most cases, until perhaps quite recently the most likely answer was - I don't. Programming statistics is something that has often been avoided by the use of "stats packages" like SPSS (now owned by IBM), SAS or more specialized packages such as Genstat.
If you wanted to program something really special then the chances are you just worked in whatever general purpose language seemed suitable or used a general purpose maths package such as Mathematica or Maple.
Then along came R.
R is an open source programming language targeted at statistics and what could be better in a world claiming to be set on creating and using domain specific languages.
As anyone with any logic would guess R was a developed from the S programing language invented at Bell labs back in 1975. It is an open source (GNU) implementation created by Ross Ihaka and Robert Gentleman at the university of Auckland (New Zealand). R comes complete with lots of basic statistical operations as built in commands but its strength comes from the thousands of plugins available to perform larger scale and more specialized analyses. It uses a command line interface for most operations but there are graphical interfaces for beginners and teaching situations. It also features lots of charts and graphs which most of the time look tend to look "academic" rather than glossy PR shots. Currently it is claimed that R has more than 2 million users.
R has been, and continues to be, actively developed as a pure open source project and it now has a commercial sponsor. In the open source world some times a commercial sponsor is needed to make an application to be taken seriously by business users. They need the promise of support and all of the things they can get from a company doing the job for money rather than love.
The company in question is now called Revolution Analytics, having recently changed it from Revolution Computing, has to be taken seriously as it is led by Norman Nie, co-founder of SPSS - one of the leading stats packages that R hopes to take users from. They aim to augment the open source offering not just by offering support, but by standardizing the graphical user interface and by improving the scalability of the core computational engine.The core engine is targeting Terabyte data sets in an effort to make R capable of the sort of large scale analysis required by commercial concerns.
The new features will not be open source and will be sold under licence.
Revolution Analytics is going some way to head off complains from the open source community by giving away the full single-user version of R Enterprise, usually $2000, - but only to academics. It has also launched a new website for the R community inside-R.org.