This heading feels useless to me.

@vondruch I am not opposed to removing it, I was thinking about someone scanning the index at the top of this very long page and trying to let them know there was easy to digest advice.

80 characters is a pretty useless definition in an UTF-8 world (in UTF-8 you have pre-composed and de-composed symbols, some codepoints have zero-width, and anyway most codepoints take more than one byte).

You want 80 columns, not 80 characters or bytes. And then most of the languages are barely getting to the point where they can count codepoints, not bytes, let alone distinguish between codepoints with or without width.

https://src.fedoraproject.org/rpms/redhat-rpm-config/blob/master/f/common.lua#_159 uses utf8.len() to compute line widths, it seems to be the best thing available today, pretty sure than in 10 years it will be replaced by something more accurate.

updated - additional comments welcome.

Commit 9331ee8c fixes this pull-request