wayne <wayne(_at_)midwestcs(_dot_)com> wrote:
{apologies for the delay}
For example, you say that 2% of TXT records are used for SPF. This
may well be true, but at the spot in the domain hierarchy where SPF
records are likely to be checked (the email domain names), I found 19%
email domain names as in '2nd level domain names' or 'owners of MX RRs'?
There's still some hierarchy in the DNS and as an example I found a single
zone with nearly 300 SPF RRs "v=spf1 -all" without an MX RR at the same owner.
Can you give me a pointer to this survey that you are referring to?
It's based on the raw data collected during the RIPE DNS hostcount for the
DE toplevel domain.
Can you explain the differences between the 2% you quote and the 36% I
found?
The hostcount works by examining every zone at every level it can get hold of.
Let's take further discussion of the survey mechanics off list, please.
My point was that neither are TXT RRs mostly used for SPF nor is the
deployment broad enough to give a pressing need to stick with TXT RRs.
Dealing with transition in the spec is OK, but that should be done in an
appendix or an informational document.
-Peter