On Thu, Sep 24, 2015 at 05:14:04PM -0400, John C Klensin wrote:
Speaking as someone who, in this area, is far more qualified to
be a victim than an expert, can we please get things that
affects what people can and cannot reasonably expect written
down somewhere, perhaps as an update to RFC 2181?
On this general issue, note that there was a recent IETF LC (and IESG
approval) of a terminology document that is only a _start_ on this.
But it is a start. We need more help.
Not to put too fine a point on it, the DNS is everywhere, and it has
evolved quite a bit. There's no question that there are gaps in the
specifications. But the last time we attempted to do document updates
without a lot of sponsorship (which is how I got to be the lucky
winner of the second chair in DNSEXT), the effort foundered for lack
of humans to review, test, and so on.
If this effort is needed, the community needs to do it. I am not
(happily) in a position to force that to happen.
Expectations about case preservation under different scenarios,
with comments about how strong those expectations should be
given various implementations
In my opinion, all you can rely upon is that case-insensitive matching
happens for any octet that happens to match the ASCII range of
letters. Preservation will happen _somewhere_ in the DNS message
(maybe in the question section only) for any conforming
implementation, but that's about it. In a distributed-management and
distributed-cache database like the DNS, I confess that I find
expectations much beyond that to be poor candidates for probable