Was this the first-ever .uk domain?
Last night I attended a party, held by Nominet at the swanky Somerset House in London, celebrating 25 years of .uk.
During opening remarks, chief executive Lesley Cowley said that Nominet still hasn’t tracked down the first-ever registered .uk domain name. I reported on this for The Register a couple of weeks ago.
After doing a little digging, I think I may have a strong contender.
ucl.ac.uk
This is the domain for University College London. There are a few reasons to believe ucl.ac.uk could lay claim to be the “first” .uk domain.
It’s well known that the .uk namespace predates Nominet by over a decade. Before Nominet was formed, registrations were handled by a Naming Committee.
According to the Milton Mueller book “Ruling The Root“, and various other sources, the .uk top-level domain was originally delegated to UCL’s Andrew McDowell. This probably happened in 1984.
Digging through some old mailing list archives, I’ve found McDowell making references to running ucl.ac.uk and cambridge.ac.uk, albeit on a test basis, as early as June 1985.
The namedroppers mailing list back then was used by academics to test their newfangled domain name system, so it’s a good place to look for firsts. I’ve mentioned it before, in this blog’s inaugural post.
In one message sent to namedroppers on June 24, 1985, McDowell writes about running .uk, ucl.ac.uk and cambridge.ac.uk on his test-only name servers. The email was sent from a .arpa address.
On July 4, 1985, he sent his first email to the list from a ucl.ac.uk email address, which suggests that the domain was up and running at that time.
That’s 20 days before .uk was delegated, according to the official IANA record.
For this reason I think ucl.ac.uk may have a strong claim to be the first .uk domain.
However, it’s possible the reality may be rather less exciting (yes, even less exciting than something already not particularly exciting).
Anonymous Coward comments posted on The Register are perhaps not the most reliable source of information, but this guy seems to know what he’s talking about:
I believe .uk was the third top level domain to be established after .edu and .us. This predated dns and would have been in 1982 or 3.
.uk was run with a hosts.txt file and the first sub-domains being either ucl or mod.
dns came in in 85 or 86 and the first sub-domains in that were copied from the UK NRS from the X.25 world (ac.uk, co.uk and mod.uk) so there probably wasn’t a first dns sub-domain for uk.
This work was done by UCL CS and at least 2 people directly involved are still there.
If that is to be believed, it looks like there may have been a “first batch” of .uk names that were put into the DNS, rather than a single domain name.
However, given that UCL was managing the system at the time, I’d hazard a guess that ucl.ac.uk was probably the first to be used.
I’ve got years worth of namedroppers posts zipped on my computer, but I can’t for the life of me remember where I found them.
In the absence of links, here’s some relevant posts.
—
24-Jun-85 03:50:35-PDT,1453;000000000001
Received: from ucl-cs by SRI-NIC.ARPA with TCP; Mon 24 Jun 85 03:50:05-PDT
Date: Mon, 24 Jun 85 11:52:25 BST
From: Andrew McDowell
To: namedroppers@sri-nic.arpa
Subject: test nameserver at UCL
We’re testing two DARPA nameservers (i.e. we’re not registered yet), and we’d like people out there to test against us. Can you reach us? does the data we’ve fed our nameserver seem sensible? and so on. From your point of view, you might want to check that you can handle the delay you’ll get. IP round-trip times between here and the states are about 2.5-3 seconds. we’re getting about 3-3.5 seconds delay for remote querys. We’re also unusual in that we have our own top level domain- UK.
We have two nameservers, ns1.cs.ucl.ac.uk. on 128.16.5.2 and ns2.cs.ucl.ac.uk. on 128.16.6.1. Ns1 (BSD4.2VAX-UNIX running BIND) should be up all day except that it reboots itself in the small hours of the morning (our time). Ns2 (BSD4.2SUN-UNIX, also BIND) may be down, especially overnight (our time). Most address records are *.cs.ucl.ac.uk, except that we also have a machine called cs.ucl.ac.uk, (giving us a mailbox on it for robert.cs.ucl.ac.uk., who happens to have a pc, also named robert.cs.ucl.ac.uk).
We have a lot of MF type records, establishing us as a mail forwarder for most of the UK ( check out your buffer lengths on interactive-computing-facility.engineering.cambridge.ac.uk.).
—
25-Jun-85 14:47:53-PDT,955;000000000001
Received: from USC-ISID.ARPA by SRI-NIC.ARPA with TCP; Tue 25 Jun 85 14:47:07-PDT
Date: 25 Jun 1985 17:48:21 EDT
From: MILLS@USC-ISID.ARPA
Subject: Re: test nameserver at UCL
To: mcdowell@UCL-CS.ARPA, namedroppers@SRI-NIC.ARPA
cc: MILLS@USC-ISID.ARPA
In response to the message sent Mon, 24 Jun 85 11:52:25 BST from mcdowell@ucl-cs.arpa
Andrew,
Our fuzzballs have joy of your domain-name servers beginning at the region cs.ucl.ac.uk, but nothing other than the question record for domains in that path closer to the root. I think your assumption is that somebody else runs those servers. Presumably, at least the top-level uk domain servers are known to the NIC. Personally, I would much rather get told by a mid-level server to look elsewhere, even back up the tree, than get told nothing. The response time across SATNET, by the way, is very good and comparable to that from NIC, at least from our swamp.
—
28-Jun-85 01:12:27-PDT,563;000000000001
Received: from ucl-cs by SRI-NIC.ARPA with TCP; Fri 28 Jun 85 01:11:32-PDT
Date: Fri, 28 Jun 85 8:28:51 BST
From: Andrew Mcdowell
To: MILLS@USC-ISID.ARPA
cc: mcdowell@UCL-CS.ARPA, namedroppers@SRI-NIC.ARPA
Subject: Re: test nameserver at UCL
Thanks for your test and message. Actually, I think we do claim to be the authoratative nameserver for the UK domain, and had NS and SOA records for UK., UK.AC., and UK.CO., but not for UCL.AC.UK. or CS.UCL.AC.UK. I’ve put these in now. Does that correct things?
—
4-Jul-85 02:03:22-PDT,2058;000000000001
Return-Path:
Received: from ucl-cs by SRI-NIC.ARPA with TCP; Thu 4 Jul 85 02:02:42-PDT
To: namedroppers@SRI-NIC.ARPA
Subject: Nic nameserver
Date: 04 Jul 85 09:59:35 BST (Thu)
From: mcdowell@cs.ucl.ac.uk
Happening to note one day that our nameserver could no longer find the address of a random site at MIT, I decided to interogate the NIC nameserver (10.0.0.51). (The other root nameserver the NIC owned up to, 10.3.0.52 didn’t seem to be returning any querys to us at the time). A request for that site yielded no address, but instead a list of names of MIT nameservers ( but no addresses). A direct request for MIT nameservers yielded those same names – but again no addresses. At last I shamelessly asked it directly about the addresses of each of these four nameservers, but it continued to avoid the question. I give up – what’s going on?
very interesting
thanks very much
i remember my first time ‘online’
it was 88/89
i was at college and was on college library pc. one had internet access
the url was christ knows how long and more dots and forward slashes than can remember but i do remember it ended .ac.uk
the screen was flickery and illegible green type and didn’t want to go online for another 8 years, unfortunately!
remember first discussing the internet in computer studies in school, 84