[arabic-vip] Draft Definitions Doc. ver 2.0 05Sep11 ..

[Andrew Sullivan]

On Tue, Sep 06, 2011 at 07:00:21PM +0200, Manal Ismail wrote:

> Thanks .. It's now more clear how registration simply makes a label
> unavailable whereas the delegation has to do with the process of the
> DNS parent pointing to the DNS of a child.  But if the definition of
> delegation is "In a DNS context, the act of entering parent-side NS
> (nameserver) records in a zone, thereby creating a subordinate
> namespace with its own SOA (start of authority) record." would this
> differ from a definition of activation? Does activation involve
> anything more ?

It could involve something more.

Suppose you have two domains, tld1 and tld2.  Tld1 and tld2 are both
allocated to TLD-corp, because tld2 somehow qualifies as a variant of
tld1.  

Tld1 is delegated: the root gets an NS record in it that points to
functioning authoritative servers for tld1.

There are now three possibilities for tld2.  The first is to do
nothing.  It doesn't work.  In that case, it is clearly not activated.

The second is to delegate it the same way as tld1 was.  In this case,
it is activated and it is delegated.

The third is to put a DNAME record into the root zone: 

    tld2 IN DNAME tld1

In this case, tld2 is _not_ delegated, because there's no DNS
delegation point.  But it is active: every query that comes in for
example.tld2 will instead be redirected to example.tld1.

Is that clearer?


>  Thanks Andrew .. In a broader context, the issue is that we need to
> know whether terms such as: register, allocate, activate, delegate,
> reserve (in variants context), reserve (in a global public list),
> block, bundle, .... etc are different, overlap or identical just to
> make sure whether we need to define any term that is not already
> defined or whether some of those could be used interchangeably
> .. Siavash has kindly volunteered to compile an exhaustive list of
> such terms afterwhich we can have an elaborate discussion to ensure
> we have a common understanding of each ..

Yes, this was exactly the set of issues that troubled us when we were
compiling that initial terms list, for the same reason.  Certainly
some of them are different to one another (e.g. see above).


> For example, if the definition of Allocation is "In a DNS context,
> the first step on the way to Delegation. A registry (the parent
> side) is managing a zone. The registry makes an administrative
> association between a string and some entity that requests the
> string, making the string a label inside the zone, and a candidate
> for delegation. Allocation does not affect the DNS itself at all."
> would a definition of a reserved variant be different?

It might or might not be.  It would appear that a reserved variant is
indeed allocated: it is associated administratively with some entity
that requests the string, although that request could be indirect
(e.g. it could happen because a string the entity _did_ request
happens to generate this variant also).

Best,

A


-- 
Andrew Sullivan
ajs at anvilwalrusden.com