If 42% of the comments on a pull request are related to increasing the understandability of the code, can we assume that just understanding the proposed code changes consumes the majority of the time spent on reviewing a pull request?