Summary

  • IANA lists .sj and .bv as Active and Delegated with Norid A/S named as manager. Those fields describe root-zone delegation, not whether a single name can be registered.
  • Only .bv has a decision on file: on 21 June 2016 the Ministry of Transport and Communications concluded that a dispensation Norid needed for a proposed Dutch-market use should not be granted. For .sj the reviewed record offers a policy sentence and a promise to evaluate, naming no decider, no legal footing and no consequence.

Delegated is not the same as open

The root database is where most people stop. The .sj record lists the code as a country-code top-level domain, Status Active, Technical Status Delegated, and puts Norid A/S in the Manager field; Hilde Thunem appears as administrative contact, Jarle Greipsland as technical contact, and the record shows a last update of 2022-01-14 (.sj record). The .bv record carries the same four names and labels, with a last update of 2026-05-22 (.bv record).

Read those fields as an answer to the question 'is this code in the root?' and they are correct. Read them as an answer to 'can I register one?' and they are silent. Neither page states that registrations are available, and the registry's own text says the opposite for both.

The identical sentence, and the identical dates

Norid's page for .sj states that the top-level domain has never been opened for domain name registrations. It tells applicants they can use .no names instead, naming svalbard.no and jan-mayen.no as examples. It commits Norid to a careful evaluation and consequence analysis before any policy change, and it says the registry will not give away or sell the code (.sj policy page). The page displays a first publication date of 2014-01-13 and a last update of 2019-06-17.

The .bv page makes the same three moves: never opened, careful evaluation and consequence analysis before any change, no giving away and no sale (.bv policy page). It displays the same first publication date and the same last update — 2014-01-13 and 2019-06-17.

Two codes, one sentence, one pair of dates. Norid's general About page keeps the status current, saying the registry has delegated responsibility for .bv and .sj and that they have not so far been opened for registration; the same page states that Norid develops domain-name policy according to societal needs and is supervised by the Norwegian Communications Authority, and it shows a last update of 2026-05-19 (About Norid).

Bouvet Island has a file

The .bv story has dates in it. Norid's chronology records that the registry began investigating possible use of .bv in March 2012, that the Ministry of Transport and Communications decided on 21 June 2016 that the requested dispensation should not be granted, and that Norid and SIDN then ended their collaboration and Norid withdrew the application (chronology).

The October 2013 assessment behind that effort is worth reading for what it treated as relevant. It weighed global use, use as a Norwegian identifier and use by a specific group; it recorded that Bouvet Island has no local Internet community because it has no permanent settlement; it asked whether a .bv model would create value for Norwegian society, be sustainable, preserve data quality and avoid serving only defensive registrations; and it explored a model limited to companies registered as BV in the Netherlands as a possible concept, not an adopted policy (2013 assessment).

The proposal that followed was not a private arrangement between two registries. Norid's own account says the proposed use required a dispensation from parts of the Norwegian Domain Regulations, that on 21 June 2016 the ministry advised the dispensation should not be granted, and that the collaboration with SIDN would therefore be concluded; the same account notes that .bv had been unused since its creation in 1997, that there was no sustainable direct use for Norwegian users, and that Bouvet Island has no local Internet community (.bv project account).

What the letter actually decides

The primary instrument is a one-page letter dated 2016-06-21. It concerns Norid's request for dispensation from specified provisions of the Norwegian Domain Regulations in connection with developing .bv with SIDN. It records that Nkom may grant a dispensation when special considerations warrant it under section 11, and it concludes that Nkom should not grant the request — in the letter's Norwegian, "ikke bør gi dispensasjon fra domeneforskriften" (ministry letter).

That is a real decision, with a decider, a legal footing and a consequence, and it is narrow. It answers one dispensation request tied to one project. It does not codify a standing test for keeping .bv closed, it does not speak to any other proposal, and nothing in it says the question is settled for good. The value is not that it proves closure is correct. The value is that an outsider can see who decided, on what authority, in what terms, and what happened next.

Svalbard gets a sentence

For .sj the reviewed record has no counterpart. There is a policy page, a refusal to sell or give away, a practical alternative under .no, and a promise that any change would come after careful evaluation and consequence analysis. There is no named decider, no statute or regulation cited as the footing, no criteria, no consequence, and no date on which the question was last examined.

Norid's administrative-model document sharpens the gap without filling it. It says the framework covers all three Norwegian top-level domains, but that because .sj and .bv have not been opened it has so far been implemented only for .no; it states that the registry holds formal responsibility for setting the .no policy and describes a framework set by Norwegian authorities, supervision by Nkom and a dispute-resolution body for .no (administrative model). Supervision exists. A dispute route exists. Both are described around the code that is open.

A promise to evaluate a future change is not an explanation of the current one. It tells you what would happen before the registry moved; it does not tell you who decided that .sj stays shut, under which authority, weighed against what.

A date that arrives before its own decision

One detail is a recordkeeping fact rather than an argument. The chronology page displays a first publication date of 2016-06-15 — six days before the 21 June 2016 decision it recounts. The same page shows a last update of 2023-04-13. A page can be redated by migration or by draft handling, and the reviewed record does not explain this one. Report it as displayed: the date shown is earlier than the event it describes.

What an outsider can reasonably ask for

The reviewed record does not say what approvals a future opening of .sj or .bv would need, so treat that as unestablished rather than settled either way. What the record does support is a narrower and more useful request. For each unopened code: the operative decision and its date; the decision-maker; the authority relied on; the criteria applied; the consequence analysis, for keeping the code closed as well as for opening it; the review date or trigger; and the route by which an interested party can have the question examined again.

Bouvet Island's segment of that file exists in outline. Svalbard's does not, in the documents reviewed here.

Why the pair matters

Norway's registry is not accused of squatting on two codes. Both are delegated to a manager that openly states it will neither sell nor give them away, and a small namespace with no registrants and no local Internet community may genuinely be better left alone. The argument is about what an unexplained status costs the people outside the registry.

Every reader who finds .sj or .bv in the root database and sees Active and Delegated has been told something true and been left to infer something false. The registry's page corrects the inference. For .bv, the file goes further and shows a ministry refusing a specific proposal on a specific date. For .sj, the correction is the only thing on the page, and the page has not been substantively revisited since 2019.

That is the asymmetry worth naming. A closure that names who decided is an instrument, and instruments can be argued with, re-tested and, if circumstances change, revised. A closure that names a policy is a habit with a date on it. Both may be prudent. Only one of them can be held to account by a party who is not the registry.

Norid A/S is the manager of record for both codes (directory entry). BTW's earlier reporting on this registry followed .no's operating surface — delegation, protocol, directory access, signing and continuity planning. This question sits one level up, at the point where a live delegation and an unavailable namespace are allowed to look identical to everyone who is not the operator.