Re: wrt: draft-ietf-dnsop-inaddr-required-00.txt

[Jerry Scharf <scharf@vix.com>]

I've been trying to stay out of this, but I just can't help myself.

This draft misses the point completely! The goal is not to require PTR 
records, since as it is pointed out, "xxx" is a perfectly valid value 
according to the draft and gives no useful information. Also, despite the 
title, there is not one MUST in the document so there are no requirements set 
forth.

IMO, for this draft to have value it must take on the issue of what is 
meaningful in PTR records. That would be worth a BCP, and pointing at it might 
produce useful benefit to the Internet. Text reenforcing the importance of 
general use of PTRs would be an appropriate part of such a BCP, but is not 
useful on it's own.

To do this, the draft must briefly examine the uses of PTR records and make 
some recommendations that find a balance between narrow and broad. It could 
even make references to things like LOC, that could also be useful if 
generally deployed. This follows nicely the questions that the chair posed.

I also believe that there needs to be a paragraph in the security section of 
any draft that reenforces the need to not blindly accept the information given 
in the PTR record.

jerry