Seriously, it's not obvious to me that it's *impossible* to change.
Of course it's not impossible. But the question is whether the benefit from the change is large enough that the people who'd have to write the software will do so.
IPv4 DNSBLs have been working just dandy with A records for a decade, and the benefits of a new RR are largely (perhaps entirely) hypothetical. It's hard for me to see anything the IETF could possibly say that would persuade the people who actually have to write and maintain the DNSBL servers and MTAs and filter packages that they have to do a whole lot of extra work when they could just add IPv6 names to the existing IPv4 names and be done with it.
R's, John PS:
could be some way of querying the DNSBL, perhaps via a top-level TXT or SOA record, to indicate whether a particular DNSBL supported the new RR's or not.
Um, wasn't the point of this suggestion to stop overloading existing RRs with new semantics just because they happen to be under certain names?
Without agreeing that a new RR is a good idea, it is nonetheless obvious how one would test whether a DNSBL served a new RR type, at least to anyone who'd actually read the DNSBL draft.
_______________________________________________ Ietf@xxxxxxxx https://www.ietf.org/mailman/listinfo/ietf