[Idr] Re: WG LC for draft-ietf-idr-deprecate-as-set-confed-set-14 (7/8 to 7/ - call continues from 7/8 to 7/26/2024

[Jeffrey Haas <[e-mail]>]

On Tue, Jul 23, 2024 at 12:07:46PM -0700, Randy Bush wrote:
> i guess i missed some discussion, but could someone please explain the
> use of SHOULD as opposed to MUST here?

Discussed here.

https://mailarchive.ietf.org/arch/msg/idr/Gz7HXS9IXBYDHVhq7AjrceGSAk4/

: > Regardless of the text in an RFC, the transition won't be immediate.  I
: > don't agree with the use of "SHOULD NOT" with the explicit intent "of
: > allowing some transition time".  Using this language provides a hole that
: > allows conformant implementations to continue to use AS_SETs or
: > AS_CONFED_SET indefinitely.
: 
: This generally overlaps the thread started on wgchairs on MUST/SHOULD
: regarding conformance.  
: 
: If an implementation is capable of generating sets in spite of otherwise
: supporting every other single thing in this document, it becomes
: non-conformant.  I.e., simple configuration otherwise does not generate
: as-sets.
: 
: Implementors and vendors should carefully consider whether they have an
: interest in supporting an RFC that they immediately become non-conformant
: with if they have no intention of immediately neutering existing code. I
: suggest you consult your own employer's product management for impact on
: RFPs.  "Partially compliant" vs. security related standards tends to
: downscore you.
: 
: Beyond that point, I'm supportive of the text above moving away from the
: "transition time" portion of the argument.  We're past that.  As I predicted
: at the microphone the first time this was presented at IDR before I was an
: author, this has mostly been a self-solving problem.

-- Jeff