On Fri, 3 Jul 2015, Liviu Andronic wrote:
While not developed much these days, Sweave has the merit of being stable and mostly ironed out; updating your installation is unlikely to break your documents.
Makes sense.
knitr on the other hand can be a moving target, and not always backwards compatible; sometimes updating knitr may end up in your documents failing to compile with obscure errors due to subtle changes. (Most of the times issues can be worked around or fixed easily.)
I suspect this will become less of an issue in the future.
This said, I would agree with Scott that knitr is preferable when you start out with R and all this knitting and weaving stuff, as it is incredibly flexible and feature-rich, and is clearly becoming the standard for the future. Soon only old dinosaurs will still be weaving their documents, as opposed to knitting them...
This ol' dinosaur looks toward the future. This is the first document where knitting the R code (visible or not) makes good business sense. I'm also developing an environmental data analysis short course and will modify the book to use knitr. Solved the problem of the function not evaluating. I put the function code in a chunk and discovered that I need to put the function call in a separate chunk. When I did that the function returned the expected value. Thanks to you and Scott for the valuable assistance. I think I'm on the way now. Happy weekend all! Rich
