[Bioc-devel] How to choose object-oriented programming paradigm for R?

[Kenneth Chu]

Hello,

Our lab is starting the process of re-implementing our existing R packages in the object-oriented programming paradigm.

Are S4 classes/methods the "best" object-oriented programming paradigm for R?

Are there other comparable paradigms that one should consider?

For example, John Chambers' OOP package, which seems simpler, but does it provide less functionality than S4?  I also noticed that OOP is apparently not available through CRAN.

Are these other paradigm, if any, as widely used, well-tested, stable and/or actively maintained as S4, and do they easily integrate with the rest of the software development facilities of R (e.g. R CMD check)?

Thank you very much!

Ken Chu, Ph.D.
Scientific Associate
Ontario Institute for Cancer Research
Toronto, Ontario, Canada