From: Russ Allbery (rra@stanford.edu)
Date: Tue Jun 05 2001 - 17:00:18 CDT
Henry Spencer <henry@spsystems.net> writes:
> That's a very strict interpretation of "existing practice". Few
> standards committees read it that narrowly. The interest in "existing
> practice" is generally centered on questions of whether the idea works
> and whether it's useful, not on what the exact syntax should be.
It's the definition of existing practice that I've seen used when one goes
into a project with the mission of "document existing practice." I agree
that it's not the definition of existing practice that one wants to use if
the goal is to "standardize existing practice" or if one is looking to
existing practice to see if an idea is feasible.
The definition of the term is to a large degree dependent on the context.
But even with the broader context, the more subtle point that I was making
is that taking a secondary header and turning it into an attribute on a
brand-new header breaks all archiving software that expects to find that
information in a secondary header, thus breaking existing practice, and I
think we need some good reason to do that in this standard other than "it
looks prettier" before including it.
(I also gave one possible reason, namely that using secondary headers
isn't always appropriate.)
-- Russ Allbery (rra@stanford.edu) <http://www.eyrie.org/~eagle/>