Well, the number of lines do matter. The original was done in Fortran. The Fortran program was the original model. It’s since been refactored into R, Python, and C++ from what I understand.
After all this refactoring, all the estimates have been changed.
A program written in this manner is almost impossible to change. Using Copy/Paste instead of loops is one issue.
Instead of subroutines just copy paste code.
Programs of this size have almost no structure either. I know for a fact because I deal with them and have been a developer for almost 40 years.
Make a small change and everything goes out the window.
Yes, size does matter.
My point is not about mistakes from not re validating migrated code.
My point is code is an implementation of a thought process, an algorithm. If the algorithm is faulty, doesnt matter how many lines or what language its in.