On Tue, Jan 24, 2012 at 10:59:44PM -0800, Murray S. Kucherawy wrote:
In short, I also find the current presentation a bit awkward. To fix it, I
believe Section 2 should be dropped, and the references for the definitions
of AA and AD should be moved to the current Section 4. No details are lost,
the document becomes simpler, and simpler is better.
Rather than moving to Section 4, could we rename section 2, adding
some explanatory text as to why it's _not_ changing the interpretation
of those bits, and then make the statements that are already there?
The reason it's structured this way is because actual fielded
implementations had mistakes around these things. We wanted to make
clear that the spec isn't wrong. (In the DNS, it's not always good to
trust the RFCs. It turns out that quite often, despite what one might
like, you get uninteroperable behaviour when you follow the RFC, and
interoperable behaviour when you look at the implementations. Yes,
this is a Bad Thing. No, I don't know what to do about it.)
I do understand the bafflement this section is causing as it stands,
however, because it seems not to be doing anything to these bits at
all. That's actually the clarification, however, and so I want to
make it explicit.
Andrew (as document shepherd)
Ietf mailing list