On 7/12/2018 3:09 AM, Dick Franks wrote:
So there's now text in attrleaf that explains about hierarchy, top,
highest, and the original presentation convention of right, but
noting that other presentations are possible.
IMO unnecessary.
This will inevitably either overlap or conflict with the draft
RFC7719-bis DNS terminology document.
I don't understand what 'overlap' you think will exist, but am pretty
sure I don't agree.
Better to use already battle-hardened terminology throughout and add
RFC7719-bis citation.
If it is that battle-hardened for this type of use, then there is no
doubt a single term in the draft that has already gained widespread use.
Which one is it?
It then declares the term 'global' as referring to the node name of
interest and only uses that term in the rest of the document.
"global" does not tick the right box for me.
And yet that's the distinguishing name of the attrleaf table in the
drafts and has been for quite a long time. There haven't been any
objections to that term until now.
Perhaps the underscore-prefixed label (sequence? / tree?) needs to be
described as subordinate to (or rooted at?) a "principal name".
Perhaps you have some usability data that demonstrates pragmatic
superiority of a particular choice over 'global' for /this/ kind of use
and can point to the entry in the bis document that already defines it?
Note that the choice echoes the use of 'global dns' that /is/ listed, to
get at the semantics of the 'reach' for the highest-level underscore name.
(Well, there are a couple of places where 'highest' was needed as
clarification.)
Stephane: "more/most general"
Except that that has no obvious semantic merit, whereas 'highest' is
directly motivated by referring to position in a hierarchy.
otherwise: "closer/closest to the root"
Why?
d/
--
Dave Crocker
Brandenburg InternetWorking
bbiw.net
_______________________________________________
DNSOP mailing list
DNSOP@ietf.org
https://www.ietf.org/mailman/listinfo/dnsop
