[ietf-dkim] Issue 1576: Revise wildcard discussion

[Frank Ellermann]

Eliot Lear wrote:
 
> The authors have chosen the DKIM style of using _adsp.domain,
> which effectively provides for subtyping.  Do you not believe
> that is sufficient?

It's fine for *ordinary* ADSP records, but issue #1576 is about
the somewhat odd case of *wildcard* ADSP records.   

>> A simple "MUST start with 'dkim='" (or similar) could fix it.
 
> But to what end? 

To the end of figuring out which of several *wildcard* TXT
records is about ADSP.

> This is where I have been bashing my head.

Maybe you missed the point *where* the wildcards are used, it's
not at _adsp._domainkey.example.com, where they would do nothing
useful, as you said.

If they are at all used it's directly at the domain in parallel
to an existing MX wildcard (or A or AAAA).  E.g. in parallel to
the existing wildcard *.claranet.de MX record.

That has already a wildcard TXT record beginning with v=spf1,
or it had that when I last checked it.  The draft (ssp-04)
says that you cannot add an ADSP wildcard, because a q=txt
for say _adsp._domainkey.xyzzy.claranet.de would then return
two TXT records, and ADSP does not know which of the two is
about ADSP.  SPF has no trouble to find its v=spf1 in this
case, SPF is only lost if two TXT records begin with v=spf1.

 Frank