For example from TFA:
"Fix code example in the field_name method"
gives:
"This commit shall fix code example in the field_name method".
OK, cool, I may or may not merge it, pull it, whatever-it. But I know what it'll do should I use it.
Now from what TFA recommends:
"Default reader batch size is 1000"
Means nothing. Tells nothing. Is 1000 good or bad? No clue. Is it causing a bug? A performance issue? Was it 1000 before the commit or after applying it? Zero information.
I'll pass and keep using the "best practice".
Using the best practices the commit would read:
"Change default reader batch size to 500" or "Change default reader batch size to 1000" or maybe "Add default reader batch size" (btw the commit line in TFA is so bad that I'm not sure at all it's 500 or 1000 or something else I should put here to make my point).
The "best practices" aren't a bias. They're the result of people thinking long and hard as to how to make commit lines as clear as possible.
I look at how multi-million lines codebase like Linux or Emacs are doing it and use that as the authority. If it works well enough for these projects, it works well enough for smaller projects.