On 2008-03-07 12:34, Dave Crocker wrote:
Sam Hartman wrote:
Making it a BCP will make the interpretation problem worse not better.
To some extent that depends on how carefully the putative BCP
is crafted, with "should" and when to disregard "should" being
very precise. What I think we've seen, with 2026 over the years,
and apparently this year with 3777, is that it's virtually
impossible to write precise procedural text that deals with
completely unexpected circumstances. Yet if the text has the
force of a BCP, it becomes very hard to interpret it flexibly
when flexibility is clearly needed. I don't know if that
is Sam's point, of course.
IETF mailing list