[Last-Call] Re: [Emailcore] Re: Re: draft-ietf-emailcore-as-28 ietf last call Secdir review

[Nico Williams <[e-mail]>]

On Thu, Apr 30, 2026 at 12:59:37AM -0500, Pete Resnick wrote:
> On 30 Apr 2026, at 0:24, Nico Williams wrote:
> > On Thu, Apr 30, 2026 at 02:42:35PM +1000, Martin Thomson wrote:
> > > The text we're debating is about the IETF telling people what
> > > configuration options they need to put into the software they build.
> > > That's out of bounds.
> > 
> > It most assuredly is not "out of bounds".  We write about options and
> > local policy all the time in Internet RFCs.
> 
> In principle, I actually agree with Martin on this point. Our technical
> specifications really ought not be telling people how to build their
> software. A careful read of RFC 2026 sections 3.2 and 3.3 and comparing that
> with section 6 of 2119 gives a good idea of how we really should be
> arranging our documents, conformance statements, and uses of imperative
> language, and maybe even described how we did 30 years ago.

Our technical specifications include protocols, APIs, and things that
are neither.  None of them tell implementors how to architect, design,
or code their implementations.  A requirement for configuration knobs is
really not "telling people how to build their software."

But I asked Martin the same question I'm going to ask you now: can you
point to a BCP that says this?  Because if not can we stop saying "X is
out of bounds" and "I agree that X is out of bounds" and focus on the
subject matter?  And perhaps submit an I-D that purports to become the
BCP that says "X is out of bounds" and let's see if we can get consensus
for that.

Nico
--