6. Review of Experimental Results

The key difference between this set of protocols and the experimental set that preceded them [RFC5335] [RFC5336] [RFC5337] [RFC5504] [RFC5721] [RFC5738] [RFC5825] is that the earlier group provided a mechanism for in-transit downgrading of messages (described in detail in RFC 5504). That mechanism permitted, and essentially required, that each non-ASCII address be accompanied by an all-ASCII equivalent. That, in turn, raised security concerns associated with pairing of addresses that could not be authenticated. It also introduced the first incompatible change to Internet mail addressing in many years, raising concerns about interoperability issues if the new address forms "leaked" into legacy email implementations. After examining experience with the earlier, experimental, predecessors of these specifications, the working group that produced them concluded that the advantages of in-transit downgrading, were it feasible operationally, would be significant enough to overcome those concerns.

That turned out not to be the case, with interoperability problems among initial implementations. Prior to starting on the work that led to this set of specifications, the WG concluded that the combination of requirements and long-term implications of that earlier model were too complex to be satisfactory and that work should move ahead without it.

The other significant change to the protocols themselves is that the SMTPUTF8 keyword is now required as an SMTP client announcement if the extension is needed; in the experimental version, only the server announcement that an extended envelope and/or content were permitted was necessary.