NMControl's handling of "ns" field makes more sense than the spec

[JeremyRand]

The .bit spec says about "ns":

"Note that this delegates all IP related responsibility of this domain and its sub-domains to the master server, effectively bypassing other settings (e.g. ip)."

However, NMControl does not delegate subdomains unless the "map" field is used with "ns" to delegate subdomains.

In my opinion, the spec is broken. There are cases where a domain owner may want to have a 2nd-level domain use a nameserver, but still have control over a 3rd-level domain, which the spec makes impossible. I do not think this is a bug in NMControl; the spec is wrong, and the spec should be revised to match NMControl's behavior.

The difference between spec behavior and NMControl behavior has resulted in some confusion among end users, both on IRC and on Reddit. See comment by @brianmay in https://github.com/namecoin/nmcontrol/issues/21 for an example of user difficulties caused by this.

Do other developers agree that the NMControl behavior is reasonable?

--- Want to back this issue? **[Post a bounty on it!](https://www.bountysource.com/issues/2845987-nmcontrol-s-handling-of-ns-field-makes-more-sense-than-the-spec?utm_campaign=plugin&utm_content=tracker%2F435873&utm_medium=issues&utm_source=github)** We accept bounties via [Bountysource](https://www.bountysource.com/?utm_campaign=plugin&utm_content=tracker%2F435873&utm_medium=issues&utm_source=github).

[JeremyRand]

The .bit spec is consistent with how DNS works; since consistency with DNS is important I suggest that we close this and open a new issue for making NMControl follow the spec. I will do so if no NACKs are received in the next week.

[brianmay]

So if I understand you correctly, you are saying that my first entry, {"ns": "ns.example.bit"} should work, and NMControl needs to be modified to make this work?

[JeremyRand]

Yes. And I already have it resolved in an internal beta.

Sent from my Android device with K-9 Mail. Please excuse my brevity.