Question: writing functions vs. line-by-line interpretation in an R workflow