Re: [ietf-smtp] Make username optional in email addresses

["Peter J. Holzer" <[e-mail]>]

On 2023-02-17 22:44:03 -0500, Valdis Klētnieks wrote:
> On Fri, 17 Feb 2023 22:40:05 +0000, Paul Smith said:
> > In the same way that Web browsers will "autocomplete" a URL of
> > "example.com" to "https://www.example.com", a compliant email client could
> > "autocomplete" an email address of "example.com" to "[e-mail]"
> 
> And how does the client know that it should auto-complete  and it isn't looking
> at a local username that happens to have a period in it?

If it's in a mailto: URI, it is simple: That needs to be globally
usable and not just locally, so mailto:example.com must be a shorthand
for mailto:[e-mail], not maito:example.com@somedomain.

(Same for other situations where a complete email address is expected)

If the user is typing the name, it is ambiguous. But here you are in an
interactive situation. The MUA could for example do a DNS lookup,
consult the local address book and maybe even perform some heuristics to
choose the most likely option and still let the user override it.


> (A quick look at this mail's From: line should convince you that local parts with a
> dot in them do in fact exist. :)

Yeah, the format firstname.lastname is very common.

        hp

PS: I don't really see the value of the proposition. Like many other
people who have their own domain (and run their own mail server) I
already have lots of different mail addresses for different purposes.
Being able to specify just hjp.at instead of [e-mail] (or [e-mail], if
I wanted to be ultra-short) doesn't make my life simpler. On the
contrary - it would cause a lot of confusion if I gave out something
which just doesn't look like an email address.

-- 
   _  | Peter J. Holzer    | Story must make more sense than reality.
|_|_) |                    |
| |   | [e-mail]         |    -- Charles Stross, "Creative writing
__/   | http://www.hjp.at/ |       challenge!"