R is a language and environment for statistical computing and graphics. Many things have changed since 1.0. The R language has acquired namespaces, exception handling constructs, formal methods and classes, much improved garbage collection, generalized I/O via connection objects, and considerable improvements in the graphics area. The user workspace has been reorganized, and so has the set of packages that ship with R. Several “recommended packages” deemed indispensable in a statistical system are bundled.

A question from laymen, how does R compare to IBM’s Data Explorer or OpenDX (in term of flexibility and graphing ability)?

R is pretty much standard if you’re a statistics major. in fact, R is the open source version of a commercial stat package called S-plus. Thus, one should think R as a statistics program rather than a data visualization program. (sorry, this didn’t answer ur question since i never used openDX)

I see R’s real competitor to be SAS (in addition to S-Plus). In the pharmaceutical arena, SAS is King, but I have noticed more and more people toying with R.

Anyone have any insight into how they compare?

I am not a big R fan (I use Mathematica), but R and SAS have very little in common. SAS is a database / data processing application that can sift through terabytes of data. It has roots in statistics, and has a statistics package included, but statistics and mathematical programming are secondary features.

I will say R is the most developed open source mathematical language. A previous poster is correct that R is more or less an open source version of S-Plus. However, it is more developed than the open source version of Matlab, Octave. None of these programs are good for data processing. Rather, they are better used for programming custom estimators and statistical tests.

R is Free. After that, R is an interpreted functional language with C-like syntax and Lisp Scheme semantics (like Dylan?) while (I’ve been told that) SAS is simply a macro processor. I got the impression that R is much more elegant as a programming language, but SAS has a richer library (in general, in several fields it has not) and that it is faster.

What R needs is an official GUI that makes shallow analyses easy for non-statisticians (a la SPSS). Personally, I’m interested with R’s interfaces to Python and Lisp.

R was intended to be like S-PLUS, but now has alife of its own.

IBM DataExplorer is not very good – its is merely a small set of tree inducers or pie-chart builders with some inflexible scope for intermediate value calculation, essentially its a light GUI over their main product, IBM DB2 database. I know, I’ve used it for a large project and resorted to either custom coding or R to do anything interesting. Co-plots with Dat Explorer? forget it. But IBM’s marketing needed to fill that gap – some data explorer type package.

Octave aims to be more like Matlab. I’d consider Scilab as it is very well featured, stable and has lots of community support. And its performance is not bad at all. I find it much more matrix-based than R which is a little more list-oriented.

To the poster who compared R to Mathematica – you are comparing different fruit. R is not a symbolic / algebraic computation system. Mathematica is. Consider the free MuPAD which again is very stable, mature and has been used by many for years. It is now linked to Scilab as I think the same academic institution (infria?) I recommend the “demos” that are provided with scilab – they will showcase much of scilab’s capabilities.

One of my favourites is Scigraphica – which aimed to be like Microcal Origin – a data presentation package … and it was very good but development tailed off. I use it to produce production quality graphs. And it is very simple yet very flexible and powerful.

All of the above (except perhaps mupad) are commonly packaged for most unix-like systems.