> They all follow the basic standards, with [] being character classes, () being submatches

You've already described a feature which has different syntax in one of the primary regex dialects I use (Emacs).