Charles Lindsey wrote:
Indeed, but "invalid DNS names" means, er, "invalid DNS names", and
"other similar uses" should cover most other applications.
No. Printed examples place no load on DNS, testing has negligible
impact. Widespread use in messaging would place a large load on DNS.
And if you read further down in RFC 2606 (I don't have the exact words
in front of me), you will see that it is recommended for use
I do have the text, and there is no such "recommend[ation]".
Well, I might be inclined to prefer using a fixed, guaranteed black-hole
valid address if somebody wants to set one up...
It doesn't need setting up. The TLD ".invalid" already has the required
property
Wrong. There is a huge difference between NXDOMAIN and an MX record that
points to a server that discards messages.
which can be aborted at once without any DNS search at all?
Gack. NOOOOO! Please tell me you didn't just say what I think you said!
In such a case, the only reasonable behavior is to query DNS for an MX
record for "REMOVE-THIS.foo.com.invalid." [*].
No, I meant exactly what I said. If it is guaranteed, by IANA
registtration, that ".invalid" will never resolve, then agents need not
bother to try it.
You have missed the point. Evidently you haven't looked at RFC
3696.
[...]
then you are just pissing
in the wind.
"Pissing in the wind" as you put it, has the property of only
affecting a small number of people. What you are proposing is
"pissing on the DNS", which has an undesirable effect on the\Internet
infrastructure.