[ih] .UK vs .GB

[Brian E Carpenter]

On 14/04/2018 14:44, Eric Gade wrote:
> Sorry one more message about this, as I've dug up an old copy of my
> dissertation (which in part addresses this issue). In 1984 when the DNS RFC
> was being drafted (I was mistaken it was the summer of 1984 not 1985), the
> proposal to use the ISO-3166 list for ccTLDs came pretty late in the game.
> There are limited records of discussions about this outside of draft RFCs
> kept by Feinler and emails that appear on the Namedroppers list.
> 
> As I said previously, they were already using .UK as their top-level in the
> NRS hierarchy in the UK. The document record shows that even as the DNS
> went into use, there was a lot of pressure for the Joint Network Team to
> revise how they structured their names, not just because of the UK/GB
> issue, but also because of the reversal of the names (hence the 1990 fiasco
> of adding Czech). The JNT had very clear reasons for not making the
> changes: they did not think the DNS would be the world standard, and had
> invested both time and money in making ISO standards a reality. They did
> not want to spend the money or time on making such changes to the NRS just
> for DNS purposes, when they believed a new international standard was
> inevitable and would require yet more changes. The X.400/X.500 systems were
> simply going to replace everything anyway, so the wisdom went.

Yes, it is reasonable to argue that .GB was just a small part of the
roadkill on the information superhighway, along with the various OSI
standards. But in my memory, the fact that there was at that time
(and until 1998) ongoing political violence in Northern Ireland meant
that the political inaccuracy of using "GB" was considered to be a
real and genuine red flag; whether you called the academic community
ac.uk or uk.ac was a side issue. JANET's commitment to OSI was a factor,
but I don't think it was the main reason for insisting on .UK.

To this day, www.qub.ac.gb would I think be problematic, since
Belfast is still not in Great Britain.

If you can track down Willie Black, maybe he could add some actual facts.
Peter Kirstein could probably tell you what went on behind the scenes.
I think he's still reachable at P.Kirstein at cs.ucl.ac.uk
(or possibly P.Kirstein at uk.ac.ucl.cs ;-).

https://wikivividly.com/wiki/.uk#History is of some interest.
https://wikivividly.com/wiki/.gb seems to know what it's talking about.

On the anecdotal side, not only was cs.ucl.ac.uk problematic;
we had the inverse problem at CERN, when supported out of Oracle's
UK office, with mail to user at uk.oracle.com bouncing after going
through a heuristic in our mail gateway. Somebody had to add a
heuristic inside the heuristic.

   Brian
 
> On Fri, Apr 13, 2018 at 10:24 PM, Eric Gade <eric.gade at gmail.com> wrote:
> 
>> The NIC collection at CHM has info about this. The draft RFCs where Postel
>> first proposed TLDs (between Jan and May 85 I believe) all proposed UK as
>> examples. Discussion on the "Namedroppers" list at the time made it pretty
>> clear why: the UCL nodes used the NRS (Name Recocognition Scheme) and
>> already had UK at the top level (though reversed)
>>
>> On Fri, Apr 13, 2018, 22:12 Patrik Fältström <paf at frobbit.se> wrote:
>>
>>> On 13 Apr 2018, at 15:12, John Levine wrote:
>>>
>>>> In another list someone was wondering why British domain names are
>>> mostly in .UK even though the ISO 3166 code has always been .GB.
>>>>
>>>> I know this came up before but can't find the discussion.  Pointers or
>>> rehash welcome.  The first mention I can find of .UK is in an example in
>>> RFC 821 in 1982, the first statement that ccTLDS would be ISO 3166 codes
>>> was in 1984.
>>>
>>> The only explanation I got orally was that "GB stands for Great Britain,
>>> while UK stands for United Kingdom of Great Britain and the Northern
>>> Ireland".
>>>
>>> That was enough for me. Don't even remember who explained it, but it was
>>> around the famous entry of .CS into the root zone that created the
>>> "interesting" situation with CS.BERKELEY.EDU (and others) and massive
>>> weird extra hacking in sendmail.cf due to the Janet "reverse" order of
>>> labels in a domain name.
>>>
>>> But, I might also have constructed this story in my head... :-)
>>>
>>>    paf
>>> _______
>>> internet-history mailing list
>>> internet-history at postel.org
>>> http://mailman.postel.org/mailman/listinfo/internet-history
>>> Contact list-owner at postel.org for assistance.
>>>
>>
> 
> 
> 
> 
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.
>