[Info-vax] bizarre SMTP problem

[Arne Vajhøj]

David J Dachtera wrote:
> Arne Vajhøj wrote:
>> David J Dachtera wrote:
>>> Steven Underwood wrote:
>>>> "Bob Eager" <rde42 at spamcop.net> wrote in message
>>>> news:176uZD2KcidF-pn2-PXhqF33pn2BN at rikki.tavi.co.uk...
>>>>> On Thu, 9 Apr 2009 19:57:15 UTC, JF Mezei
>>>>> <jfmezei.spamnot at vaxination.ca> wrote:
>>>>> Then it appears to be a bug, since '+' is perfectly valid as a character
>>>>> in an email address, as per RFC 2822 section 3.4.1.
>>>> That may be, but there are lots of systems that do not honor plussed
>>>> addressing.  My mail provider (SpamCop.net) allows for plussed addressing
>>>> but many web based systems find those addresses to be illegally formatted.
>>> Such characters are not valid in UN*X usernames for many flavors of
>>> UN*X, nor would it be allowed on VMS. So, it seems likely that an
>>> attempt to verify the existence of such a user on a VMS or UN*X
>>> receiving end would be likely to fail.
>>>
>>> The protocol itself my not disallow it, but is that a valid username on
>>> the receiving system, and if so, how is that validated?
>> The part before the @ is not necessarily a username.
> 
> Well, yes and no.
> 
> Technically, it's a mailbox name and there is - USUALLY - a strong, if
> not direct, correlation between the mailbox name and the username.
> 
>> Today it is
>> practically never a username.
> 
> That is inconsistent with my experience. In my experience, the mailbox
> name is almost always the user name, unless the messaging system
> provides a correlation between the mailbox name and a username at some
> level.
> 
>> 15 years ago it did not have to be
>> a username (PMDF supported firstname.lastname back then).
> 
> ...and even today it does not NEED to be a username, but typically is,
> especially in the Windows/Exchange world. 

My experience is that the majority of corporate email use
firstname.lastname as preferred email address instead of
username. And that practically no ISP's use username
for email.

Arne