Re: Effect of “should certainly do foo” in policy
Ben Finney <firstname.lastname@example.org> writes:
> When Debian policy says “should”, there's no mandated minimum severity,
> but it generally merits a bug report.
> What does policy mean when the “should” is intensified with “certainly”?
> Examples include:
> 3.4.1. The single line synopsis
> The single line synopsis should be kept brief - certainly under 80
> 10.4. Scripts
> Shell scripts (`sh' and `bash') should almost certainly start with
> `set -e' so that errors are detected. […]
> Do violations of “should certainly” merit a bug report? How does this
> compare with an unmodified “should”? How does it compare with a “must”?
Severity-wise, it's the same as "should." Policy is not written in formal
standards language, and certainly here has the normal fuzzy English
meaning. In the first case, for instance, I'd read it as an attempt to
further explain what "brief" means, acknowledging room for disagremeent
short of 80 characters. For the second, the "almost certainly" is an
intensifier on the certainty of the recommendation, but it's still a
recommendation rather than a requirement.
Anything about the one-line synopsis is, in practice, probably a minor
severity bug even if it's over 80 columns, just because that's the bug
severity used for documentation problems and things that are ugly but
don't break the package.
> If there's no normative effect, is it reasonable to ask for the sake of
> clarity that these modifiers be struck from the wording?
If the information they're communicating can be rephrased in some other
way, sure. (In other words, I disagree with simply removing the words.)
The ideal solution would come as part of a general Policy rewrite to use
more formal and precise language.
Russ Allbery (email@example.com) <http://www.eyrie.org/~eagle/>