On Apr 15, 2016, at 11:16 AM, Sean Leonard <dev+ietf(_at_)seantek(_dot_)com>
wrote:
On 4/15/2016 10:23 AM, Arnt Gulbrandsen wrote:
John C Klensin writes:
Unless people propose to update RFC 5321 to eliminate a
requirement that has been in place from 821 and through 1123 and
2821, I don't see that there is anything to discuss. It seems
to me that the rules are very clear, i.e., that, except on the
final delivery SMTP server, two mailboxes are equal iff:
-- The domain parts are equal under DNS rules
(case-independent for ASCII strings and U-label:A-label
equivalence for IDNA strings)
-- The local parts are equal if they are octet-by-octet
identical.
If you ask people to type in their address in a web form, the addresses will
largely be typed in by people whose email addresses are case-insensitive,
and many of them know it. Rather like their names and street addresses,
which have proper casing but not essential casing.
You're suggesting that although the user may know the address to be
case-sensitive, the software used should absolutely not consider it
case-insensitive.
At the end of the day, I do not see a problem with the status-quo in the mail
standards, i.e., "case-preserving: DO NOT MESS WITH IT".
If the user types fooBAR(_at_)example(_dot_)com, take that as input and don't
change it. If it delivers, great. If it doesn't deliver, ask the user for
something else that delivers.
It should not be any intervening software's job to change the user's
/expressed preference/, except for the delivery MTA. Whether the delivery MTA
delivers foobar@ and fooBAR@ to the same place or different placeds, that is
the delivery MTA's choice, alone.
Why again is this so complicated?
As an example, if I go to https://www.ietf.org/mailman/listinfo/ietf-smtp and
unsubscribe the email address "STEVE(_at_)wordtothewise(_dot_)com" that'll
remove
the email address "steve(_at_)wordtothesie(_dot_)com" from this list, but if I
try and unsubscribe
"steve-ietf(_at_)wordtothewise(_dot_)com", it won't.
Those three email addresses are all distinct, but they all deliver to the same
mailbox. The mailing list server - which is not the delivery MTA - understands
some sorts of canonicalization, and not others.
Any system which uses email addresses as identifiers needs to understand
that to some extent if it wants to provide a decent UI to users who aren't RFC
lawyers.
Cheers,
Steve
If you ask people to type in their address in a web form, the addresses will
largely be typed in by people whose email addresses are case-insensitive,
<<
*Are* email addresses case-insensitive? This is not an essential property of
an email address. You, as an implementer, user, sysadmin, can treat strings
as case-insensitive or case-sensitive as you wish. But your choice does not
impact how the delivery MTA treats it. Just because you treat a bag of bits
as "your property" doesn't mean it becomes your property. Only the delivery
MTA has a say in the matter, that matters for the primary purpose of an email
address: to deliver electronic mail.
Rather like their names and street addresses, which have proper casing but
not essential casing.
<<
True of some nations and postal systems, but not all. Only the national
postal system gets a final say in the matter. Ask China vs. the United States
vs. Turkey: they have different rules for what characters will result in
delivery to the same physical mailbox.
Sean
_______________________________________________
ietf-smtp mailing list
ietf-smtp(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf-smtp
_______________________________________________
ietf-smtp mailing list
ietf-smtp(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf-smtp