In the last dplyr vignette "rowwise" the author did note about rowwise performance. For example, rowSums is faster.
Actually, I don't know a good place to suggest so I leave it here.
As I know, dplyr is much more suitable to be column-wise and a lot of rowwise operations will be handy in coding and typing but slow.
Are there any plans on passing matrix operations into rowwise?
There are tons of str_ operations. What if create row_ operations like row_sum with a search of predefined optimized operations by first letters in RStudio?