Re: [dane] Encoding local parts in better ways

[Sean Leonard <[e-mail]>]

I read draft-levine-dns-mailbox-01 with interest and have the following 
editorial comments (and one technical comment):

Section 1:
Many systems allow "extensions" such as john-ext or mary+ext

The term used in the vernacular is "sub-addressing" (in everything that 
I have seen), and the term in [RFC5233] is "subaddressing" (no hyphen) 
or "detailed addressing" (with space). "Plus addressing" (occasionally 
"plus-addressing") and "minus addressing" or "'minus' addressing" has 
also been seen in the wild. I would change that sentence to say that:

Many systems allow for subaddressing, such as "john-ext" or "mary+ext", 
where "john" or "mary" is treated as the effective local-part, and "ext" 
is passed as detail information to the recipient for further handling. 
See [RFC5233].


Section 4.1:
and synthesizes an key record,

change to:
and synthesizes a key record,


Technical:
Abstract
"and the DNS design that only does exact matching."

I thought that DNS uses ASCII labels, which are matched 
case-insensitively, and that there are no other modern options. 
[RFC4343] [RFC6891] False? See also Sections 3 and 4 in the I-D.

Sean

On 9/20/2015 8:54 PM, John Levine wrote:
> I've sent in a new version of draft-levine-dns-mailbox-01 that
> describes a bunch of ways to encode mail address local parts in ways
> that don't need canonicalization or address guessing.
>
> Take a particular look at section 5, which publishes regular
> expressions to match a domain's mail addresses.
>
> * Can represent any plausible local part syntax including case
> folding, noise characters, multiple ways to write Unicode characters,
> suffixes where some are ignored and some aren't, BATV, and VERP.
>
> * Reasonably fast lookup (max of one query per localpart character)
>
> * Works fine with static zones served by ordinary name servers .
>
> * Doesn't make bulk addresss harvesting easy.
>
> If you really want to do experiments in publishing mail info in the
> DNS, I think this would be a rather interesting one.
>
> R's,
> John