Some years ago, I started doing self code-reviews for most of my PRs.  It
allows in-line elaborations of the changes (that are *not* better put as
code comments), and it sometimes reveals little accidents / things I
overlooked and need to follow-up on.  It helps other human reviewers
understand / trust what was done, which ultimately leads to better / more
efficient peer review.  I review lots of PRs, and I appreciate this a lot
when I see it!  It seems more common now.

In the current age of AI with total/mixed/unknown provenance of a PR, I
think this practice is even more important.  It helps demonstrate to other
reviewers that the human author/contributor has taken responsibility for
what they are presenting for anything non-trivial by reviewing it
themselves.  When I don't see a self review, and I see some questionable
things, I begin to lose trust and wonder "did you even look at this" and
feel my reviewing time was taken for granted / wasted.

This may not have presented a problem here yet; I'm not motivated by an
experience in this community.  Well not recently, anyway.  Nonetheless,
given the benefits of self code reviews, I'd like to edit our PR template
to *recommend* that a self-review be done immediately afterwards,
particualrly for anything non-obvious.

Any thoughts on this?

~ David Smiley
Apache Lucene/Solr Search Developer
http://www.linkedin.com/in/davidwsmiley

Reply via email to