First Contact (Internet at 50) – Computerphile

First Contact (Internet at 50) – Computerphile


So I came to Nottingham in 1980 to do a physics degree and Worked through that but towards the end of it I got very interested in computing and fact, I ended up doing physics with computing that’s where I met Dave and the other people who were setting up the early computer science group and Then he said after I’d finished that I’ve got a bit of spare money. Would you like to work on some typesetting stuff? There’s all sorts of interesting things happening So I did. I worked on that for a year, I think, with Dave doing various things including this connection for the exams machine and all that interesting stuff Never mind said Julian quite rightly. So, we’ll keep the lower level And then I got really interested in networking us as that was just starting to come about and I worked with another lecturer called Hugh Smith and he had contacts with UCL and so we got into this sort of Internet and ARPANET and Early connections to that and I started working on email software to try and make sense of all this stuff that was going on get rid of all these exclamation marks in your routes and things like that and Just try and make it simpler It was quite early on so it was about 1984 or three I think I got involved in it. At that time The the internet was well, it was all just ARPANET. So it was the American connection and we had a very very tenuous connection we had some friends at University College London, and they had a satellite link to the ARPANET they’d had for a while and we were in the maths building which is where we are today and We didn’t actually have a connection to the outside world But (the) Psychology (department) did. So you could connect across to Psychology’s computer and dial up London Log into one of London’s nodes and then you could use the internet Very very slow. And of course it was going across satellite So you connect to… I don’t know, there weren’t very many sites, but there were a few places you can pick up software So you’d log into those? but because it was going over a satellite you typed a letter a and it would take about three seconds for the a to come back and so You’d start typing and then you get very confused halfway through and have to delete it all and this was all painfully slow But we thought it was wonderful at the time So it was about you know, 20 minutes or something to get onto the Internet from here But but we did manage it. Alongside that at the same time There’s sort of a more anarchic network called Usenet, which was just – that was machines That you were friendly with. You talk to other machines and and a whole grouping came out of that so we also have that so you could use that alongside it so you could Send email to somebody and say do you know a good site I can go to to pick up software for? this, that, or the other and they’d send something back over that but it was all very very disconnected so, I remember we connected to Lancaster and then Lancaster connected to Kent and Kent connected to Amsterdam and Amsterdam was the European gateway. They could then connect – and er, at a very early stage you actually have to work out your route so if you wanted to email somebody you’d say right so go from Nottingham to Lancaster to Kent to Amsterdam to Seismo to Rutgers to something other and you could eventually work out where to go and how it would get there So it was So much fun We had email in the department for a long time and then as psychology and other places got connected. We sent it around campus. I can’t think who I would’ve send the first one to probably somebody at University College London because we did a lot of work with them early on and eventually had some Connection which is how we got connected to the Internet I do remember desperately trying to get in contact with Brian Kernighan and Sending email after email along these different paths – via Amsterdam and Seismo to… It was called “Research” – Research Everybody’s Name was just their computer name So “Research” was the AT&T computer node, so and eventually I remember it was late on a Friday night I was watching the logs and suddenly saw an email come in. It was from Brian Rushed down to Dave Brailsford, and said look, look, we’ve got a connection here’s the route that you need to use and he scribbled it down and From then on we had contact there very soon after that. We started sort of collaborating. So I was working with somebody in the California and we were working on software that got switched around so I’d be working on it during UK time and then he’d be working on at Californian time and then he’d send me changes and Yeah, that was mostly done over email we didn’t need Very much Because in the UK we were still we weren’t using tcp/ip. We were Told we had to use x.25 at the time But we cheated and we ran some tcp over x.25 and we ran strange protocols that allowed us to Connect to the Internet and we did actually even managed to have a sort of talk System where you could type a message and it would appear in California and vice versa That was a great day when that first happened I guess that was about nineteen eighty six or seven when that we managed to cobble together enough protocol to get that all to work SEAN>And you mention email the instant thing you think of these days? It’s probably webmail or a client email with attachments just being dealt with for you. Can you talk me though that? You couldn’t do attachments at all though, you could sometimes because it was all just text so sometimes you could wrap them up in interesting ways and There was particularly one self-extracting It was called a shar archive self-extracting it was actually a small shell script program for UNIX and so Your message arrived and you just ran the contents of the message as a program and it extracted itself and you could send files that way but That was fairly late on because we had very little bandwidth at the time. So it was more very quick short emails and This is a place to pick up a file from we would try and get the file somewhere near nearer to the person So that they could pick it up with a reasonable speed. So this is a message I’ve saved from 1987 and this is the path you can see here So it starts off over here. This is the person and then this is all the places It’s been through, Seismo was a popular place in I think it’s in Virginia or somewhere like that But it had great connectivity to lots of other places and then down here We have MC VAX, which is the European node And then it trickles across the UK to get to us from here So, you know That was kind of one of the early messages and you had to remember these paths and write them down so that you you could Get back to them and this is kind of the route it took it jumped around in Europe a little and then left over across the Atlantic and then rattled around there until it ended up where it went to and this is just a, This is the Sat net that was So this bit here is all of the US and there’s a connection here coming over the satellite to Goonhilly Down in Cornwall and a landline up to UCL I remember when they finally got to upgrading that they went to a transatlantic cable or something TAT 8 or TAT 6 Transatlantic telephone cable 6 when that was laid down. Suddenly. We had more bandwidth than we ever knew what to do with Until about two months later when we fully utilized it all That’s the way it goes That was what a lot of the internet was for and that’s still what the Internet’s for actually, serving very little purpose but you know back then they took it to a to a different leveland we’d have to go to this really dark room and The computers were green littleI had an Amiga computer, right? This is like an Amiga 500

57 thoughts on “First Contact (Internet at 50) – Computerphile

  1. Any chance of an episode on FidoNet?
    I found the following in a Fido nodelist from 1990 on MIT's website, it looks like you guys were running a node!

    2:254/414 Ian Sherman Nottingham UK Nottingham University

  2. I was just reading a RFC document for OAuth which is RFC6000 something, and I wondered which was RFC 1. I opened it and saw it was published in 1969 and thought "holy shit". Then this video came out.

  3. There is an excellent book on the men and difficulties involved on the creation of the Internet, first published in 1996, "Where Wizards Stay Up Late: The Origins Of The Internet". Available in hard and soft cover, digital and audio book versions. It talks about the people and what they went through. Written in simple terms and the few places where it gets technical, written so almost anyone can understand it.

  4. My first email address (with exclamation marks) didn't go through seismo, it went through ihnp4. Then there were the % addresses, usually to go from one network to another "user%[email protected]"

    It almost makes me feel old.

  5. This made me really curious about how the internet developed in the early days. I guess initially there were small local networks that were connected with each other over time? It would be extremely interesting to see a visualisation of all the connections at the time and how new ones were added…

  6. This is SOOOOO ridiculously interesting! I knew a decent amount of the story (not from Nottingham’s point of view but in general) but there's just some kind of charm that shrouds the early ARPA/internet today for me as a 16 year old. And as a blind person, I miss the days when almost everything on the internet was text. May have looked ugly but it was definitely easier to navigate with a screen reader.

  7. This video is too quiet. It's difficult to hear what is being said unless I blast the volume on my computer. If you right click on the video and select 'Stats for nerds', you can see that YouTube measures the volume as 11 dB too quiet. It would be great if you could double check the volume on future videos before you publish them.

  8. The ARPAnet was a single WAN (wide area network) which transmitted its first packet in 1969. The Internet is, as its name says , a connection between networks and its first three network packet (one of those networks was the ARPAnet) was transmitted in 1977. From then on the same ARPAnet hardware moved not only its own traffic (NCP) but also increasingly Internet traffic (TCP/IP) until NCP was switched off in 1983, making the ARPAnet vanish into the Internet. The story in this video happened right after that. Just saying "ARPAnet was the early name of the Internet" is a far simpler story but is technically wrong.

  9. 5:57 Hahaha, nowadays we consider that kind of laid back attitude as a recipe for disaster! Running a shell script sent over to you by mail without examining its contents and/or being sure of what it does? That's how you get viruses 😀

  10. Ok, the joke is on me, this channel is holding a grudge against people who can't understand spoken English, because after my constant complaint about not having subs, they don't fix it. Brilliant, you people deserve a noble prize in the douchebag category.

  11. if anyone wants to play with shell archives, and pretend they are back in the '80s! Here you go!

    ftp://ftp.gnu.org/gnu/sharutils

  12. But did you forget that we had a fully fledged (X25) based network called JANET (joint academic network) that supported a full suite of applications including interactive, file transfer, job transfer, email and directory services. It also connected to Europe via EARN. In many ways more advanced than the early ARPAnet.

  13. I used Usenet and UUCP way back in 1986 or 1987. My home computer at the time (first an Atari ST and then various versions of Unix for the PC) still shows up on old Usenet maps. 'omnifars'. You needed maps for bang paths. For email you had to do source routing, basically tell the network which path through the network you wanted the email to take.

  14. I'm wondering why the psychology department had a better network connection than physics, maths, or engineering.

  15. Old people reminiscing about old computers…
    That's why I switched to math, even older people reminiscing about even older equations! 😉

  16. And us poor civilians not in academia could not get access but we had FIDONET, all done through dial-up modems that went on and off the network at haphazard moments. An email might take days to arrive if an important node wasn't switched on for some time. And we thought it was fantastic!

  17. THE COMPUTING HISTORY VIDEOS YOU HAVE ON COMPUTERPHILE ARE POSSIBLY SOME OF THE MOST VALUABLE VIDEOS ON THIS STUFF ONLINE, EVEN IF THEY CAN BE DRY TO SOME

  18. Thank you for the subtitling! And it even isn't 'just auto-CC' (wich i personally don't mind at all), but full fletched manually inserted subs!! 😀

  19. I was at UC San Diego from '81-'86, and was one of the few undergrads there working on what was intended to become the BSD4.2 release. BSD4.1 was a hit in data centers (and campuses), and the need for more inter-system protocols was urgent. These were needed in the worst way possible, and that was what BSD4.2 contained, in particular a very weak TCP/IP implementation. Slowly the consensus grew that 4.2 should be treated as a "lesson learned", and that all effort should shift to the much cleaner and more complete 4.3, which as a result was released only about 2-3 months after what would have been the (much delayed) official 4.2 release date.

    The 4.3 release was also hastened by the great need Sun Microsystems had for a better UNIX for their nascent workstations (which I used from their first generation). Sun was one of only 2 or 3 companies to try to push 4.2 into the real world, and Bill Joy himself led the charge away from 4.2. I don't recall if Sun issued grants to the BSD project itself, but I do know they paid the time of their own engineers and many interns to work on 4.3.

    4.3 also superseded the "bang notation" used to hand-specify network routes. Before 4.3 there were no agreed-to ways to discover and monitor "live" network connectivity, and it seemed everyone had their own suite of ad-hoc methods. The Ham radio community had FidoNet, which had more network awareness, which influenced (and was influenced by) USENET, another store-and-forward network.

    The majority of Internet protocols layered above TCP/IP are text-based, primarily to avoid incompatible binary encodings and "endianness" issues. This also made them much easier to debug, needing only a text dump of the session. This in turn meant a human could "speak" the protocol directly over a telnet connection, the human acting as a client to a server for that protocol. For years I used to routinely check for new email using only a terminal directly to the email server, typing the SMTP messages by hand (later a very short Expect script, then a Perl script). I'd fire up my email client only when I had unread messages. We'd do the same thing to get directory listings on FTP servers (to check for new files having landed), and to do early IRC sessions.

    More existing text protocols are being replaced by binary versions, such as HTTP/2. And other protocols have essentially become binary by being compressed in-flight. If it weren't for WireShark, I'd hardly be able to speak Internet any more!

  20. I know the university near where I live (Rutgers) gets around and has been a few interesting places in history, but I'm always surprised when I see its name appear in something like this. Fun stuff!

Leave a Reply

Your email address will not be published. Required fields are marked *