Dear WG,
Thank you for your thoughtful feedback during the WGLC for the revised
lame delegation definition. With this email, we close the WGLC for
rfc8499bis.
With the discussion and feedback during the interim and with the WGLC on
the mailing list, the chairs have determined there is rough consensus on
the proposed wording on lame delegation, and as it is now included in
the latest revision of rfc8499bis on datatracker.
After the document shepherd write-up, the document is sent to the IESG
as the next step.
Best regards,
-- Benno
On 18/07/2023 08:17, Havard Eidnes wrote:
Note that Scott's current EPP draft is still using this term,
citing the definition in 1912. Should the term be removed
from Scott's draft, or acknowledged that it is now historic?
If Scott replaces it with another more precise term, can we
get that term in this document so that future uses can cite
this document?
[SAH] My draft uses the term in only one place where it describes
practices that can "introduce risks of lame delegation". I'm
inclined to change that sentence to something like "introduce risks
of invalid DNS delegation" to avoid the term completely if it
eliminates the need for a normative reference that doesn't yet
exist.
Don't take this the wrong way -- I'm not picking on you presonally...
But I think this illustrates one of the problems with coming up with
"specific and clear" alternatives to the "lame delegation"
characterization which is also brief, precise, and doesn't have too
wide difference to normal-language usage.
Because... Even though my first language is not English, I would have
thought that it ought to be possible to distinguish between a "valid"
and an "invalid" delegation without querying for the actual current
operational state wrt. to the given zone, i.e. there has to be some
evident "rule violation" involved. Quoting the Merriam-Webster
dictionary:
invalid
: not valid:
a : being without foundation or force in fact, truth, or law
| an invalid assumption
| declared the will invalid
b : logically inconsequent
The most relevant part here would be "without foundation in ... law",
if we consider the RFCs "law" within this area, which isn't too far-
fetched.
E.g. it would IMHO be "invalid" to use an NS record with a name server
name which contains "_" in its name, because there is at least a deep-
seated convention (if not a rule from ... rfc 952?) that host names,
i.e. names which resolve to A and/or AAAA records (and therefore, by
extension, name server names), should not contain "_" in their owner
name (by default enforced by BIND to this day, when loading an
authoritative zone), i.e. "it explicitly goes against the rules".
(And, yes, I know that not everyone agrees with this rule...)
I would claim the flavour is distinctly different, and that "invalid
delegation" is not a good substitute for "lame delegation".
Regards,
- Håvard
_______________________________________________
DNSOP mailing list
DNSOP@ietf.org
https://www.ietf.org/mailman/listinfo/dnsop
_______________________________________________
DNSOP mailing list
DNSOP@ietf.org
https://www.ietf.org/mailman/listinfo/dnsop
