> what i'm urging here is great caution both for the authors and the > reviewers. improving the readability of this topic over what we wrote in > resimprove-00 seems necessary but is not a simple proposition.
I don't think this is actually all that complicated. The problem is that you are trying to give implementation advice, and you just shouldn't be doing that. Stop giving implementation device and it gets easy: Names subsidiary to a cached NXDOMAIN MAY be assumed to be nonexistent for <interval>. Some cache implementations may be able to realize performance improvements for queries under cached NXDOMAINs by making this assumption. Implementors of caches are expected to be smart enough to figure out whether this is true of their implementation. That's obviously not detailed enough, and the last point doesn't need to be stated explicitly, but that's the basic point you want to get across. Anything more than that is too much. The reason the WG is getting pushback from me on this is precisely that the draft gives implementation advice that I know is not universally applicable, and quite unnecessarily states that advice in the form of a strong normative recommendation. This is not something the WG should be doing. Implementation advice should never be even a little bit normative. _______________________________________________ DNSOP mailing list DNSOP@ietf.org https://www.ietf.org/mailman/listinfo/dnsop
