On 14 Feb 2020, at 12:52 pm, Paul Hoffman <paul.hoffman@icann.org> wrote:
On Feb 13, 2020, at 4:08 PM, Geoff Huston <gih@apnic.net> wrote:
a) why does DITL rate a special mention?
It is a term that appears in other RSSAC documents that is not widely known by all readers.
fair enough I guess - it still seems discordant to me to pick out this particular program within the scope of this document.
b) why is anycast not mentioned specifically?
We tried but failed. Different folks have very different views of how to define it. Do you feel that a definition of anycast is required in order to understand instances?
yes
If so, what part?
err - can’t parse this question. "what part” of what? Of “instances” or of “anycast” or of “”something else entirely”? Can you rephrase your question please?
c) why the use of the term "root server” used as a collective noun, whereas the industry convention is to use the term “server” in a more unitary sense to denote an instance of a operating system stack. I had heard in the past the term “root letter” to denote the collection of instances operated by a root server operator, and the use of “root server” in this context is a new one on me
This was heavily discussed on the list over the past month. So far, there is a strong agreement not to use "root letter" or anything about letters.
Even so it does seem slightly jarring to me to see technical terms that have accepted meanings in other realms be redefined in what I would say is completely different way in this context. Agreement or not, borrowing a term that means something in most other contests (a singular host of a service) to redefine it as a collective noun solely in the root server context seems to be odd. It seems to be adding subject-specific jargon rather than de-mystifying it. thanks, Geoff