Re: I-D Action:draft-newman-email-subaddr-01.txt

[der Mouse <[e-mail]>]

>> Whereas, if the MLM is explicitly told that the convention is in use
>> within the ADMD (at least for this subscriber),

This raises another hazard of a convention like this: it can lead to
things like MLMs assuming that anything with a + (or whatever
deliminiter is chosen) is an instance of this convention, leading to
severe brokenness when faced with addresses at a site that doesn't work
that way.

FWIMBW, I'd prefer to see emphasis on the "nobody but the receiver and
those with out-of-band knowledge of the receiver have any business
assuming anything about local-parts" aspect.  If my system happens to
have an address with local-part *^%+_!^! (or, more plausibly, one+two)
nobody without special knowledge of my system has any business
inferring any structure from that; in particular, inferring that
local-parts *^% (or one, respectively) even exist, much less bear any
relationship to the full form, is broken - absent, of course, specific
out-of-band reason to think the relevant mailer uses plussed addresses.

>> That said, it has been suggested that MLMs might want to go a step
>> further, and rewrite messages such that a message from
>> <[e-mail]> copied to a mailing list which has a
>> (Subaddress-enabled) subscriber of <[e-mail]> would
>> cause the MLM to go as far as rewriting headers to the subscribed
>> address.
> This makes me nervous.

Me too.  Your reason strikes me as valid, but I'm actually more worried
that implementors will lose that "(subaddress-enabled)" part and simply
assume subaddresses - after all, it's Standard, right? - and end up
sending mail to the wrong places.

/~\ The ASCII				der Mouse
\ / Ribbon Campaign
 X  Against HTML	       [e-mail]
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B