Editorial

The ‘Editorial’ article from the 2010-2 issue.

Welcome to the 2nd issue of the 2nd volume of The R Journal.

I am pleased to say that we can offer ten peer-reviewed papers this time. Many thanks go to the authors and the reviewers who ensure that our articles live up to high academic standards. The transition from R News to The R Journal is now nearly completed. We are now listed by EBSCO and the registration procedure with Thomson Reuters is well on the way. We thereby move into the framework of scientific journals and away from the grey-literature newsletter format; however, it should be stressed that R News was a fairly high-impact piece of grey literature: A cited reference search turned up around 1300 references to the just over 200 papers that were published in R News!

I am particularly happy to see the paper by Soetart et al. on differential equation solvers. In many fields of research, the natural formulation of models is via local relations at the infinitesimal level, rather than via closed form mathematical expressions, and quite often solutions rely on simplifying assumptions. My own PhD work, some 25 years ago, concerned diffusion of substances within the human eye, with the ultimate goal of measuring the state of the blood-retinal barrier. Solutions for this problem could be obtained for short timespans, if one assumed that the eye was completely spherical. Extending the solutions to accommodate more realistic models (a necessity for fitting actual experimental data) resulted in quite unwieldy formulas, and even then, did not give you the kind of modelling freedom that you really wanted to elucidate the scientific issue.

In contrast, numerical procedures could fairly easily be set up and modified to better fit reality. The main problem was that they tended to be computationally demanding. Especially for transient solutions in two or three spatial dimensions, computers simply were not fast enough in a time where numerical performance was measured in fractions of a MFLOPS (million floating point operations per second). Today, the relevant measure is GFLOPS and we should be getting much closer to practicable solutions.

However, raw computing power is not sufficient; there are non-obvious aspects of numerical analysis that should not be taken lightly, notably issues of stability and accuracy. There is a reason that numerical analysis is a scientific field in its own right.

From a statistician’s perspective, being able to fit models to actual data is of prime importance. For models with only a few parameters, you can get quite far with nonlinear regression and a good numerical solver. For ill-posed problems with functional parameters (the so-called “inverse problems”), and for stochastic differential equations, there still appears to be work to be done. Soetart et al. do not go into these issues, but I hope that their paper will be an inspiration for further work.

With this issue, in accordance with the rotation rules of the Editorial Board, I step down as Editor-in-Chief, to be succeded by Heather Turner. Heather has already played a key role in the transition from R News to The R Journal, as well as being probably the most efficient Associate Editor on the Board. The Editorial Board will be losing last year’s Editor-in-Chief, Vince Carey, who has now been on board for the full four years. We shall miss Vince, who has always been good for a precise and principled argument and in the process taught at least me several new words. We also welcome Hadley Wickham as a new Associate Editor and member of the Editorial Board.

Season’s greetings and best wishes for a happy 2011!

Note

This article is converted from a Legacy LaTeX article using the texor package. The pdf version is the official version. To report a problem with the html, refer to CONTRIBUTE on the R Journal homepage.