[ietf-dkim] Issue 1576: Revise wildcard discussion

[Eliot Lear]

Frank Ellermann wrote:
> Stephen Farrell wrote:
>
>    
>> ssp-04 did revise the wildcard text, but not exactly as suggested
>> in the issue, nor am I clear about whether the new text satisfies
>> the couple of people (Eliot, Frank) who commented in the thread.
>>      
>
> The version in ssp-04 IMO misses the following wildcard TXT points:
> (1) There is no explicitly specified way to identify an ADSP record,
>      when it comes as one of several TXT records in a q=txt reply.
>      In the terminology of an IAB draft ADSP defines no TXT subtype.
>    

The authors have chosen the DKIM style of using _adsp.domain, which 
effectively provides for subtyping.  Do you not believe that is 
sufficient?  I'll argue that the use of _adsp is actually better in that 
you don't have to parse through a bunch of crap to get to the 
appropriate record (normally).  You still need the code checks, of course.

> (2) Even if ADSP would do this a set of wildcard TXT records for
>      various purposes (compare RFCs 1464, 4406, 4408, and 3920bis)
>      might be too long for UDP.
>    

And so I don't see how this holds, given the above.

> (3) As a consequence of (1) ADSP likely doesn't work for wildcards.
>      As a consequence of (2) the WG apparently refused to fix (1).
>      A simple "MUST start with 'dkim='" (or similar) could fix it.

But to what end?  In what circumstance would a wildcard that stops at an 
existing label be at all useful?  This is where I have been bashing my head.

Eliot