Ewart Shaw (J. E. H. Shaw), University of Warwick
I am a lecturer in the Statistics Department at Warwick, and J has been my main programming language since around 1992. I find J invaluable throughout my work, both because of its power (one can try new ideas very quickly) and because of its ability to handle large projects.
J is a fantastic research tool. For example, I have for many years been developing numerical and graphical methods for Bayesian inference, such as the use of codes and related highly symmetrical objects to construct efficient high-dimensional numerical integration rules. J is ideal for trying out and implementing new ideas like this. I have also been developing a statistical modelling package called BINGO (Bayesian Inference: Numerical, Graphical and Other stuff) based largely on this work.
Apart from the benefits of the core J language itself, J can interface internally with highly developed existing software like LAPACK and interface externally with a front end, language or package with which the end-user may be more familiar.
All my lecture notes (several hundred pages) are constructed using J, including the associated input to and output from free packages like Bugs, (La)TeX and R as well as commercial systems like Mathematica and SPlus. I also use J to produce hundreds of individualised assignments and corresponding answer sheets each year.
J's handling of large projects makes administrative duties easier, and even bearable. For example, my J timetabling system deals with the 200+ different modules taken by our 500+ Statistics students, and typesets timetables for the various degree courses, years, streams, lecturers and lecture theatres.
In my spare time (Ha!) I have also used J to maintain websites, manipulate images, write music, analyse board games, and solve widely varied problems & puzzles. J is fun.