Democracy and the Internet: New Rules for New Times
Tim Berners-Lee
Europeaum Policy Forum
The Sheldonian Theatre, Oxford University
Wednesday 26th September

Well thank you. It's quite moving to be here. I'm coming back to Oxford after graduating here 25 years ago. Pity to come at such sad times. I come from Boston, where in the town I live most people know somebody who has lost somebody, if they didn't lose a friend or colleague themselves. So these are sad times.

So, the Web and democracy - I'm going to look at it generally. The way we self organise as human beings is such that we can manage our common things with self respect, with mutual respect, and like the Web it's all about communication. Now, going back not 25 but ten years to the start of the Internet, late 1989. What I really wanted in those days was to make the Web an information space built on top of the Internet. I wanted it to be a space where people could communicate by sharing information.

I don't see the process of communication as being just my saying something and your doing your best to understand it and if you look puzzled my saying it again or you asking questions. Communication is a lot of really interesting processes. We work together and then, in this room, we're exchanging things. In fact, I'm going to be as quick as I can so that we can get back to our questions. As we have this exchange, we, in fact, build up new concepts. We're not just trying to transmit the old ones. The exciting thing is that we are building new ones. By the time we leave this room this afternoon, we will be a little bit of a group. Because we have that common understanding because we have a slightly better shared use of some vocabulary, for example. We'll also have erected a little fence of thin information boundary around us.

When we use those terms amongst ourselves, people outside will have a little trouble understanding and this is always a trade off. This is always the tension. This is all about culture and sub-culture, and there has been a question about how you structure civilisation since time immemorial. So, nothing in principle has changed that. But it's a really interesting way, I feel, of looking at the problems of democracy.

I don't regard frankly the problems of e-democracy as getting everybody to be able to vote for an elected person, or getting everybody to respond to an email, which has been sent to every single person allowing them to express their opinion. A very large portion of people get enough email as it is. You certainly can't just operate as one big mass. We have a lot of choices of structure. Whenever you have a sub-culture or sub-group of any size, you make a community work better within itself. You create a very important boundary and boundaries are as important as communication.

Some people saw the Web as being a wonderful thing for global communication, for information, which is free and public, and read by everybody. It is wonderful. There is a huge amount of information, which can be read by anybody. (It doesn't have to be read by everybody, I think that's where some people get panicked, where they feel they have to read everything!

Since the Web started, I've had two questions very often, two fears people have presented to me. One of theses fears is, for example, from the French. If we open the pipe to America, then won't our culture be destroyed by McDonald's coming down the pipe from the States? Now why they think the pressure of culture from Mc Donald's should be stronger than the pressure of culture from the Louvre, I don't know. It's a serious concern that if we have a global network, we will homogenise our culture. We'll end up with the lowest common denominator. Language will contain only those concepts, that are sufficiently bland to be understandable by absolutely everybody and so we will lose all the richness. We would lose Welsh and Greek, and it's clear to me that would be horrific for a couple of reasons.

I get a horrible feeling looking back at history, at some of the attempts to create the one perfect system. Some of them have failed. Fortunately, there have been other systems which have in the end worked better and been supported better by people who cared, and they have ended up suppressing the ones that were damaging.

A homogeneous system is clearly very dangerous. We need people with diverse ways of looking at the world, different sub-cultures in the world so we don't want that homogeneity.

But at the same time, the other fear that people express to me is now we have the Internet surely we'll get these cults? They don't even have to be in the same place. People can just be in the same mental space because you can filter your email. You can only visit websites that you've been pointed to from that email, which is from people who are a little bit crazy like you, and those websites will point to each other and very soon you'll end up in this cultural pothole, which is very, very deep and very, very slippery. When you stagger out of your basement into the sunlight because you have to get some more coke and pizza, you bump into a real person another person who's not in your same cultural pothole and the only common language you have is violence. This is an abstract problem and was without a specific example until two weeks ago.

So, it is important as well that while we have diversity, there is a balance and it seems that it's not only society which clearly needs a balance. In its way, a lot of nature does. Does anybody remember fractals? Anyone know what a fractal is - a pattern - like the coastline? As you get closer, you fly above the coast and it has a certain interesting structure. Then you fly down to a tenth of the altitude and it still has an interesting structure. You get closer and closer until you're looking at the way the seaweed curls around a few of the pebbles and it still has an interesting structure. It has structure in all levels. I have a gut feeling that society needs to be like that.

In order to build a democracy, you have to have structure. It can't be a simple structure, which operates just at one level. Think of some societies which have gone wrong and you'll find that they've been much too flat. They haven't had an interesting sub-structure. We know also that a simple tree doesn't work. We need a complicated structure, which is fractal in some way. That means that our society and the technology which we use to support it has to work and pay attention at each of these levels.

So let's start at the bottom. When you're thinking about how to spend your time, maybe you should take the hours of your week and divide them into boxes, depending on whether you've been dealing with personal time or family time, which is about the next box up, national things, world issues. There are ten orders of magnitude between yourself and the world population, so you'll need ten boxes. Look at the end of the week and see how you've spread it - see whether you've got a balance in your own personal use of time. Let's quickly go through a few of these levels. Firstly, the personal level - how does the Internet at the moment support it? Concerns are rife, of course, about privacy ("privacy" I can say over here), and there are ways in which technology can allow a person to protect themselves. I think it is a worse problem in the States than it is in Europe, because there is less legislation over there.

In principle, we have technology that has come out of the World Wide Web consortium it's called P3P - the Platform for Privacy Preferences Project. As you go into a site which starts to ask you for information, your browser will automatically pick up the privacy policy of the site. It will not just pick it up as an inordinately long piece of very, very small print. It will pick up a matrix, basically a mathematical matrix. It will say for standard types of information, or standard uses, we will do the following things with this sort of information from you.

So there are ways in which, in some of these areas, technology can help. But it won't work unless there is legislative backing. For example, if a company can just totally ignore all these privacy protocols, then it only takes one rogue site to extract all this information and sell it to the highest bidder. Then the consumer/individual protection has gone. There are a huge number of issues in which the boundary around the person is at stake. But a lot of these are not new with the Internet. The Internet sometimes shines a rather harsher, brighter light on it than before.

Now let's look at how we use the Internet in larger groups; within the Consortium (by the way, I work for the World Wide Web Consortium, a sort of standards body). We have working groups which work with perhaps 12 or 20 people in a group, interacting with maybe 12 or 20 other groups trying to put together new technology. These groups try to use the technology as much as possible for collaboration. They have to collaborate across a distance - nobody likes leaving their family and friends, and flying too much. So we've tried to give each group a part of the Web, give each group a place that any member of the group can go at any time of the day or night and chat. They can see who's around to discuss issues. It turns out, there are a lot of things you need to be able to do. There are a lot of mechanics you have to put in which are tricky.

When you buy a Web Server off the shelf, you don't get the ability to control who has access to information. It is in fact something which is a very important part of the process of producing documents. People won't share things if they feel it is immediately going to be public all the time. For the group to work, it needs this information boundary. The group has to be able to control access to its crazy idea, the crazy idea that it is a bit nervous about. Then it needs to be able to determine who it is going to negotiate reviews with. Later on it comes to the horrible point where it has to expose its views to the world in general - when it realises what a deep information boundary it has built around itself.

Then, hopefully without moving all the documents and rewriting them all and changing all the URLs, it needs to just be able to give more people access. So we've found we've had to put a lot of tools in which didn't generally exist to allow control over these social things, such as who has access, who can read documents and who can write documents. We found that as we allowed more people to write them, we had to do things like keep track of every single version so that people knew they could go back in case one of their colleagues made a mess of it and so on. So in fact, you need a lot of powerful tools and a lot of things that we just still don't have to make a group work.

Now we're going to look at the way families are using the Web. For me, this is really important. A lot of people are very excited about the way they've managed to share photographs. But very often, you find that to share photographs within a family they've put them on a commercial server, which has made them public. Now some servers actually allow you to go and create a group, and you can do that for your family. After you've come back from your holidays and your family have all gone back to their individual homes, you can get together on a conference call. You would like to be able to sit in front of your screens, sharing and building a photograph album together, agreeing or laughing about which ones you'd better file out because you wouldn't want them to be seen by future generations but eventually ending up with a common view of what were the high spots of the holiday. It's like actually a design process or building laws. It's a very creative process and you're putting together a common understanding. You're building the family up as a unit and to do that you need all these tools and you need them available in a very intuitive way because you want your grandmothers to do it as well as your grandchildren and that just isn't there.

The technology needs a whole lot of more deployment. So moving up, how do we use this for larger groups? There's been a lot of talk about voting - I won't talk a lot about that. I tell you one thing I would like to see - and maybe I'm just dreaming - I would love to see a new genre of communication evolve on the web. I would like to see a debate with accountability. I've seen the sort of rhetorical debate, which maybe the Oxford Union has a lot of responsibility for, in which if you can win the point of the moment, you've won the point. But if the debate is going to be the stuff of making policy when somebody says: "Two thirds of the children in the United Kingdom are doing this" and the other says "No, oh no they're not only a third of them are", I would love to see that with links. I'd like to see the people responsible for those statements go back and justify it. I'd like to be able to go back and see a version of that. Goodness knows, if this is an important debate, it's going to make an important decision for the country, I would love to see a certain amount of responsibility and I'd like to see the incorrect statements being withdrawn eventually.

Now, of course, there are some things you can argue over till the cows come home, but there are other things that you can't. Wouldn't it be wonderful if even the television talk show is later cut into little snips and put on a website and decorated, and then the people held responsible following weeks would say: "Now first of all we'd like to read you thirteen things you said last week which were proved to be demonstrably inaccurate. Would you like to apologise before you come on the stage this week?" Dreams, dreams, dreams.

So moving up to bigger areas - the nation. I have a problem with the nation at the moment. There are a lot of flags flying in America and when I think of it as a symbol of togetherness and of the enormous amount of warmth which is being shared between all Americans at the moment, it's wonderful. When I see it as a symbol of one country against others it's pretty frightening. So when we look at the balance between these top two levels of nations against the world, obviously everyone realises we're in a critical position, where the gap, the digital divide, is just the one extra thing on the bottom of a very long list of differences between the developed and the developing world. It's just one more thing between the rich countries and the poor countries, which we need to fix.

The frightening thing here is that the world is getting smaller to the extent that you can almost put your arms metaphorically around it. The hope is that maybe we are getting to the point where the Internet and radio will get there and there will be understanding. There will be enough families exchanging snap-shots and the level of penetration will get to the point where we do have enough global harmony to prevent war. It is just something that people have been dreaming about. Maybe the Web can help do this. I think it's an open question as to how you increase the amount of communication. Will you eliminate these isolated potholes, or is there something about the way people behave that there will be certain people always who, even if you give them the ability to communicate and to see how other people live and to realise that they are actually human beings and have the same issues as you, will still fight and rebel and do damage? I don't know.

I guess it's a question of psychology. On a world scale, I think the divide is a very serious problem. I'm glad it's being addressed on the UK side. But frankly, looking at the UK and the USA, I'm not too worried because I think it will get there. Televisions have basically got there, telephones have got there, Internet will get there. I'm glad people are worrying about it because it is only if people are worrying about it that they will get there. But the bigger gap is between the developing countries. We get these awful questions. Is it really just a slap in the face with a wet fish to give someone a high-speed optical connection when they don't have clean water?

I've heard stories both ways - I've heard a story from a missionary who went in to try to help restore people's morale in a war torn area, about a boy who had somehow got himself an Internet connection. Having learned English from a translation of the Bible, he managed to set up a business doing translations into his local tongue. He was actually bringing money into the village by doing this translating service, which seems too good to be true. The money could then pay for PVC tubing to bring in the piping. So that's one argument.

The other argument is that it's ridiculous to put in the Internet before you've got basic healthcare. All I can say is we just have to put our efforts into doing all these things and we can do it out of altruism or we can do it out of selfish feelings. It's going to be equally important for both for a stable world. Talking about stability, when the crisis happened on the 11th, around where I was the cell-phones were basically unusable. The system was so overloaded because everyone wanted to call home, to call their loved ones and say that they were OK. A group of us were in a meeting at a retreat, where we deliberately didn't have a lot of connectivity but there was a phone line. Being geeks, we immediately set up a network. We had a wireless network within the room, so laptops were connected to one of the major geeks who acted as a relay and firewall, connecting through the one telephone line with a modem. This allowed everybody to send very small amounts of information and pick up the odd picture of what was going on. Then one other person went out down to a local store and bought a radio.

I think the Internet demonstrated its worth because we had a chat system running with the folks back at MIT who were taking the calls from relatives. This chat screen was projected into the meeting room but we were aware that Internet connectivity was very fragile. The Internet is famous for being a net, famous because if you cut one piece of it there are so many other routes round. However, in practice, because we base it all on the telephone infrastructure, we have made it fairly tree orientated - there are still single points of failure.

For example, I have various cables coming into my house, and several of them are capable of taking Internet traffic but I don't have a little box which will use the best one, or the cheapest one at a given time. If a tree falls across one wire, I don't find that it slowly degrades - somebody has to go and put the wire back on. If the link between my cable company and MIT breaks down, well I can't talk to MIT for a while, there are all kinds of cables which link them up indirectly to MIT but they haven't made the system so that it actually will automatically respond. So in fact, it relies on people and I feel in fact we could make it very much more resilient.

It's really important to have the Internet connected, available, always on. For many things such as e-voting - you don't need broadband, you just need the cable always connected.

The problem with dial-up is not primarily that it's very slow, it's that you have to dial - what I'd like to see Stuart Hill do is get every house in the UK connected permanently. I'm not worried about whether they can watch videos. I would just like to be able to press a button on the screen on the fridge and bring up the school menu for today - know whether I have to make a decision about whether I'm going to send the kids with money or food, and without the time it takes to dial in.

I'd like you to help me with a problem, because I've spent 10 years thinking about the social effects of the Web in various ways, but this is the next piece of technology.

It's something I call the Semantic Web. Let me explain what it's going to be and then maybe you can try to tell me what the problems will be or what the advantages will be. Before the Web in 1989, if you used computer documentation systems, if you were very lucky you could use the same PC to dial into one system and go through the library access system, and Telnet and various different programs and eventually you would get the gem of information.

You would copy it into the paste buffer and go into a totally different system using different programmes, completely different knowledge of how to do it and find a place where you want to enter the data and paste it in and save in whatever way. Very often you couldn't even do that because the data formats were completely incompatible.

So now with the Net we've got, wonder of wonders, an abstract space in which all this information seems to be the same, it seems very consistent if you can use the same program. You don't even have to worry about what programme you're really using - you have the impression that you're just in a space of information.

In fact it's all stored on weird machines and weird systems still, they're all still there -- those strange mainframes running strange programmes but they've got this little layer of software which maps them into the general Web space. This is very wonderful for human readable documents, but it's not there for data. You don't have to think of a big database. Just think of the data in your life, if you have a pocket gadget which holds your calendar, your appointments - that's data, its got things like times, and it's really fairly organised data.

And you may have a desktop machine that runs a different sort of programme but does the same sort of personal information management. You may have bought that from a different supplier and if you take these two things and you want to connect them together the state of the art at the moment is that you have to buy third party software from somebody else who makes a living selling software. It's rather unsatisfactory.

Imagine that you were going to a meeting and you found a Web page describing this meeting and you think "I want to go to that meeting". At that point, your agenda really needs to know when you're going to be there. You may have in your car or pocket a GPS device that will actually help you get there - if it knew where it was and at what time. So for that day it would be nice if that information which is on the website were in your address book. You'd cut and paste it from the Web. You'd get out one gadget and you'd copy it into that, you have to fiddle with the dials on the GPS to set the latitude and longitude to where you're going to have to be and then you cut and paste some of the stuff about the meeting. Once you've done that none of the hypertext links work so you have to save the Web page somewhere else on your desktop in case you want to follow any of the links.

This is ridiculous. When it comes to the data in our lives, we are pre-Web. What I mean to say is if you understand what that information means, all kinds of programmes will be available. Your telephone answering machine will be able to say that you're not there, knowing that you're several hundred miles away, and the possibility of connecting different applications becomes huge.

There is a huge amount of money being spent in industry for integration of enterprise applications, taking the fact that the inventory system has a database with some information and accounting software has a database with other information - in fact there ought to be an overlap but isn't. What happens is that you call in a consultant, who does this third party business of analysing how one company has stored the information and charging you a hefty fee for specialist software just for you to integrate your enterprise applications.

Think about the dates on the Web. There's a site which has the weather for example, there's a huge amount of information about the weather out there and a huge amount of information about stock prices.

Suppose I asked the simple question, could I write a programme that will use this information to find out whether the stock price of a company is related to the weather changes at its home-base?

It would take forever, of course. It would be horrific because the programme has to pretend to be a person and go to the Website, read the Web page, strip off all the pretty pictures and guess that in the third column down in the table is the temperature. You have to scrape the data out of these Web pages that are intended for people - this is ridiculous!

There are clearly defined relationships between these things. There are mathematical relationships and we should make languages that are mathematical.

Without getting too carried away about it, this is really exciting. Thinking as a frustrated software engineer, which is what I was 10 years ago, you get the impression that if we actually solve this problem then things could be very different.

So how is it actually going to change the world? Well, for example, if all the financial information of companies is available in a way that could be treated then any student can sit down and write a programme which would, for example, look for possible signs of financial instability.

The savings and loan disaster happened because they said people haven't really looked at these savings and loans institutions sufficiently well. Maybe the data was available, maybe it was pinned on a tree in a forest somewhere but maybe it was actually publicly available. Now if we start to get this information available on the Web, maybe there will be this informed electorate - they will be able to write programmes and they will be able to analyse and use more and more powerful tools.

Now this is a bit of a shift, you can imagine. It's really not very exciting to imagine what happens when you have two semantic Web documents with data and a link between them. It's just as unexciting as when I showed people the first two Web pages.

What you have to do is imagine what would happen if there were billions and billions of them out there and in particular you have the option of making a hypertext link point in principle to anything.

Same thing with the semantic Web when you realise what this person means by postal code is exactly what this person means by zip code and suddenly two huge pieces of data can be correlated. There are possibilities for huge, very, very easy abuses of personal privacy here. Possibility perhaps of doing fascinating new scientific research by correlating different pieces of things. Possibility of building in little gadgets, little robots, programmes which just wander around the semantic Web looking for correlations, just in case they find things interesting.

When the Web started, one of the things we knew was that it had to scale, and you'd never be able to make a list of everything which was on the Web. We were proved wrong. What we hadn't realised was yes, the Web was going to grow ridiculously fast, but so would search engines. Combined with a few interesting mathematical results, which make things like Google work in an uncannily effective way, that's really changed the way people use the Web and the way the Web can be used.

What will be the equivalent thing for the semantic Web? What are people going to build on top of that? Will it be that artificial intelligence suddenly comes into its own because there will be a whole lot of stuff out there which is machine processible, which doesn't need natural language interface, which is really processible?

How will we be able to use this ability to process large amounts of data? It's not something that will naturally hit people like the browsers. I won't be able to take a semantic Web browser and give somebody that great a "aha" moment by suddenly letting them browse around for a little bit. But on the other hand, being able to access data which computers can help us analyse, may in fact have more dramatic consequences in the long run.

I want the Web to be fair. I think society should be fractal, The one optimistic thing I have is that when I look at people I think that most people do actually put their marbles fairly evenly into all kinds of different pots.

There must be something that drives them not to always spend time at one particular scale. There must be something that evolution has given us so I think we're naturally disposed to behave such that society becomes fractal and everything will be all right - and so I leave you on that optimistic note.