I've been using lme4 in R for linear mixed-effects modeling for a couple of years. I have some older scripts and data, and now I need to check a few output numbers. However, when I run the analysis on the same data by using exactly the same scripts, the output numbers are all different from those I got several months ago. Sometimes it's just a slight difference, but sometimes it's a critical one. How is that possible? Could it be that the script developers made some changes in underlying math? Is there a way to get back to previous versions of calculations?