Now I've seen the new RFC 6186 specification (https://tools.ietf.org/html/rfc6186) that describe a "DNS way" to achieve the same goal.

[..] again: [..] When a user identifier is required, MUAs MUST first use the full email address provided by the user, and if that results in an authentication failure, SHOULD fall back to using the "local- part" extracted from the email address. This is in line with the guidance outlined in Section 5. If both these user identifiers result in authentication failure, the MUA SHOULD prompt the user for a valid identifier. [..]

here I have the biggest doubt, in fact Thunderbird often uses only the username and not the whole email address, both for incoming and outgoing server... Is any of you using this second way? And possibly, do you kniw if it works with different clients (desktop and mobile)?

I've made some tests and this solution seems very serviceable, for the only exception for the user name on Thunderbird. Unlike what is stated in this RFC, Thunderbird "takes" only the name and not the entire email address as the username. Anyone knows if is it possible to communicate the username type via RFC?

Many thanks again!

Davide