Re: should vs must

[Julian Gilbey <J.D.Gilbey@qmw.ac.uk>]

On Thu, Mar 01, 2001 at 01:21:27PM +1000, Anthony Towns wrote:
> > This is also a nice piece of advice, but is orthogonal to the
> > suggestion being made.
> 
> Uh, reread Sam's message: he was saying that there would be a number of
> guidelines that would always have exceptions, I was disagreeing.

An example:

   If one person maintains several packages he/she should try to avoid
   having different forms of their name and email address in different
   <tt>Maintainer</tt> fields.

There may be a number of reasons why this is the case.  We will never
say that the person MUST use identical forms, just that it is normally
the case and is recommended practice.

This says that in a perfect, bug-free distribution, there will still
be exceptions to this recommendation.

That is what Sam pointed out:

   This "should" will never be an absolute requirement, as there may
   be genuine exceptions.

   But there are other times when policy uses "should" to mean "our
   target is that every package follows rule XYZ, but we're not yet at
   a stage when we are able to demand this compliance."

The suggestion is:

   When we propose a policy change, we discuss whether we wish it to
   be an absolute requirement at some point in the future, or whether
   there will always be exceptions.  We then use MUST/SHOULD according
   to the target.  The question of whether we are ready to demand that
   packages yet follow this specific directive is a separate question.

   Julian

-- 
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

         Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London
       Debian GNU/Linux Developer,  see http://people.debian.org/~jdg
  Donate free food to the world's hungry: see http://www.thehungersite.com/