Arnt Gulbrandsen <[email protected]> wrote:
> 
> 2. The localpart is opaque, the document says.

   ... which is, architecturally, quite correct.

>    The portion to the left of the at-sign contains a string that is
>    globally opaque and is called the <local-part>.  It is to be
>    interpreted only by the entity specified by the address's domain
>    name.
> 
> That's not entirely practical. Anyone who implements address-book 
> search, "unsubscribe" handlers and code that needs to look up addresses 
> will find that one has to do lookups case-insensitively.

   None of these are architecturally correct.

   Address-book searching will indeed find different capitalization of
what seems to be the same address, but that's more a user-interface
issue of mail-readers.

   Automated unsubscribe SHOULD find a way to match something actually
sent in a list posting, rather than guess from any of several email
address fields. But in practice, anything that "works" is probably
"good enough"...

> I would like to have an additional paragraph in section 3.1, stating 
> that although the localpart may or may not be case-sensitive, as the 
> domain owner pleases, some software and users do treat it as if it were 
> case-insensitive anyway. Software MUST preserve the case, and SHOULD do 
> address lookups case-insensitively.

   I disagree with this point. We've had enough tweaking; and we're
describing architecture, not best-practice.

   "MUST preserve the case", IMHO, is already implied. "SHOULD do
address lookups case-insensitively", IMHO, doesn't belong in an
architecture document.

--
John Leslie <[email protected]>

Reply via email to