Tim Berners-Lee invented the World Wide Web. And the web, as he conceived it, was made as much of principles as protocols. It would be free, open, and permissionless. Anyone could link to anyone. And in a now-unthinkable act of good faith, Berners-Lee refused to patent his invention. Remarkably, this crazy scheme worked. The web crushed the other Internet protocols like Gopher, and also the surprisingly Facebook-like America Online and its nerdy cousins, Prodigy and Compuserve (even if it did take a decade). Now, the web is so ubiquitous that many people think the Internet is the web.
So, as life accomplishments go, Berners-Lee could have retired, victorious, at the age of 36. (In fact, Berners-Lee has been knighted for his work, so please call refer to him in your head as Sir Tim as you read this interview.) But Berners-Lee has continued working and thinking about technology and society. He's been an animating presence and conscience of the web and Internet throughout its existence. From his role as a professor at MIT, he's helped the web and the principles that informed its creation adapt to changing business and technological conditions. His most recent research focuses on how to link together data and sensors and things, not just documents with text and pictures on them.
With the rise of mobile apps that are very unwebby and growing social networks that contain an increasingly large percentage of web user's outputs, "the future is quite in the balance," he said. "Everybody is trying to take over the world. That is the commercial imperative with the capitalist system. But against it, there is the force of creativity, and the excitement of the jungle outside the walled garden." And in this battle between the forces of enclosure and those of freedom, Berners-Lee has earned the right to his optimism. "The permissionless, free-as-in-freedom web always ends up winning," he concluded.
We spoke with Berners-Lee in his office in Cambridge by Skype.
In the future, do you think there’ll still be something we call the web that will have the features that you built into the original specs?
I don’t predict. It’s hard and risky. I prefer to have a wishlist of things we ought to try to build. Right now, I think the future is quite in the balance. We may end up with a very recentralized Internet, which is controlled completely by a few large players. The industrial government complex, if you like. There are a lot of questions.
Back in the 90s, people were all concerned because Netscape had complete domination. Then they woke up on Monday morning and they were more worried about Microsoft dominating not only the browser, but the operating system as well. Then they stopped worrying about Microsoft and started worrying about Google. Then some of them said, "No, I’m not worried about Google, I’m worried about Facebook."
Often, what’s happened is that the tension has shifted to another area of the space. So, we’re less concerned about the lower levels. Yes, of course, there is still a danger that an ISP gets a complete monopoly in a given country and can abuse its users with prices or with spying and censorship and so on.
But in general, I think we’re all set with a competitive landscape in terms of providing the actual connectivity. And we’re still enabling massive creativity. While it’s a platform for everybody’s free and unfettered creativity, it’s gonna be fine. It’s gonna be a force for good.
What kind of web do you want to see in 10, 20 years? What’s on that wishlist?
I want to see one where the underlying infrastructure is open. So that nothing happens in the middle.
I want to see one where we’ve moved from worrying about getting the last 60 percent online to worrying about how to get the last 10 percent online, which will be a different issue. I want to see one where the web is affordable, where people can get online with using less 20 percent of their income, and more like 5 percent of their income.
So, it’s ubiquitous, it’s affordable, it’s open, it’s neutral. You can make a website. I can go and see it, no matter how we connect to the Internet. It is also permissionless. I didn’t need permission to invent the web. You didn’t need permission to start a website. It’s really important that whatever it is in the future continues to be permissionless.
I don’t know if you happened to read Jill Lepore’s blowout New Yorker story on link rot and our desire to archive the web versus what seems like a technological tendency towards decay and entropy. It seemed to me, for so many years, that things would stay around—that you would be able to not just address the present moment but the past through this web of linked documents. But it does feel like that assumption is crumbling because pages are changing so quickly, because people lose domains. You’re in the Lepore article, of course. Brewster Kahle says, “I’m completely in praise of what Tim Berners-Lee did, but he kept it very, very simple.” You didn’t build in this temporal element, the versioning aspect.
Well, on our own website, we did. The W3C is backed up and has a versioning system. Responsible websites do track change. My stuff at home, my stuff at work. This is not something we make available to the general public because it obviously takes some effort.
But if you look at places like Github, the history is being tracked.
I think part of [the problem] is that I used the domain name system because it was there. I remember when the domain name system became seriously commercially used, Esther Dyson was the first person in charge of ICANN and she told everyone how great it was that domains were rented, not sold.
And I thought what’s wrong with this picture! But that was the ethos. It was an exciting place, things come and go. Finding somebody else using your domain name for something else next week was part of the excitement of it.
I still think the domain name system needs to be changed. I have talked about having something like Docarchive where they are not rented, but given. And when you use one, and if you go bankrupt, someone will take it over.
But it does take a decision. As webmaster in chief at W3C, I said once we use a URI [uniform resource indicator] for something, we will never use it again for something different. The responsible thing is keeping the same URIs and it is not rocket science. It is very easy to map stuff. The URI is really a name for things, not a location.
The other really pressing problem with the web right now is that mobile browsing is basically broken, even on big professional media sites. I even started a Tumblr called Broken Mobile Web, where if something didn’t work, I’d screenshot it. And I probably had 35 examples in just a few days. Even in the 1990s, things weren’t broken like that.
Well, the 1990s web was broken in that the pages were just designed for a laptop [with a certain size screen].
So, we had this mobile web initiative, which was all about teaching people how to make mobile. There was a "Mobile OK" logo you could put on things. That’s interesting as a user that you think it’s a very frustrating experience.
Initially people said, you must set your computer to 800x600 pixels, otherwise you won’t get the full benefit.
You don’t get that option anymore.
There was a guy who wrote a book all about how to do that and then wrote another book recanting, saying, "You’ve got to make these pages as device-independent as possible."
But that was the big screens coming in that made them realize pages should be device-independent. I’ve got the 30-inch screen in front of me. And when I’m reading a big article, it would be very nice, if I could split it into columns, but very few people have thought about that optimizing for big screens because it’s rare. Maps work well.
Part of the permissibility of the web was how simple it was to make a page, but now it is radically more difficult the second you even add an image to a page because of all the screen sizes. Do you see a way out of that? Is there a way that the rendering of these pages can deal with these things more easily?
That’s a really good question. So, what tools can we have? The natural thing is to have more semantics in the web page. To say, in a blog, "This image is an important image, which you should see at full size." Responsive images work, which in a way makes life more complicated, but clearly something like blog software ought to do a good job. Because when you write a blog, someone is usually reading an article and they are looking at the occasional picture, but the software should know well how to adapt it to a phone. So, when you have a particular genre, it should be adapted very well by people putting effort into the blog software.
But when you’re writing a web page and sending it out there, it would be nice to have a high level way of saying, "I want this to go into columns when it is very big, or one column when it is very small, and I want all the margins on the side to disappear when it is very small, and these images are just decoration."
Do you see Facebook as a continuation of your thinking about the web as a medium where people consume and create—or as a competing set of ideas made into code?
Both, I suppose. Looking at any big social networking site, I won’t pick on Facebook or LinkedIn or Google Plus. To some extent, they are great websites. They get user involvement. They get user creativity on a good day. On a good day, they have URls so you can link to them.
On a bad day, they try to ensnare you. They become walled gardens. You don’t have a web experience. Within a certain network, can you tweet or like it, or are you pushed into doing things within the single walled garden?
So, two issues. One, they have a tendency to become walled gardens. And two, they have a tendency to seem like monopolies. Whenever we’ve had a monopoly, there is always a threat that it severely hampers innovation. But it’s amazing how people can switch to something else.
One of the things I’ve been tracking is that according to a Times article, Facebook has been talking to publishers, saying they want them to start publishing directly into Facebook more often, and they say it’s to fix the issues with the mobile web, because they can serve pages more efficiently and faster to mobile users. So, instead of Fusion publishing things on Fusion.net, they’d be publishing directly on Facebook or publishing a second version of that story inside the walled garden, formatted and structured just for Facebook.
What’s wrong with this picture? Certainly fragmenting the web, so that one copy is inside the walled garden, is terrible. It’s important that when people are looking at articles, they all see the same URLs, so all the discussions about it can be attached to it and shared across different places where people have discussions.
Obviously there is always a massive commercial incentive to try to take over the world. I don’t know if you’ve seen Pinky and the Brain, it was a cartoon series where Pinky says to the brain, “What are we trying to do today? Well, we’re trying to take over the world.” So, everybody is trying to take over the world. That is the commercial imperative with the capitalist system.
But against it, there is the force of creativity, and the excitement of the jungle outside the walled garden, the permissionless, free-as-in-freedom web always ends up winning.
A couple of weeks ago, W3C launched the Web of Things initiative. I want to know how you think the web will change as more actual thing-things have web addresses.
In the world I’d like, you’d be able to run programs in your own house which do lots of useful things, which are intelligent but don’t necessarily go through the cloud. They might back stuff up in the cloud, but they don’t require you to go through the cloud for everything. Even devices like door locks connect through the cloud. Hello!
I’m looking for a smartlock which I can run through Bluetooth to my phone and then my home computer, where I don’t have to connect to any cloud provider. The trust people have in cloud providers is not great anyway. Why should my lock connect to the cloud and have its security depend on some app provider?
How does our relationship with data need to change in your view?
Well, this connects to another general new direction we’re going to take for the web with my personal data: the lock on my front door, the temperature in my living room, the fitness data captured by the accelerometers on my phone, the data from my hospital from my blood results.
There is quite a large amount of data out there about me. Some of it is from sensors, but a lot of it is from other places, too, like other organizations.
But no matter where it comes from: it's my data. It’s my data! I want to keep it where I can control it. So, I am looking for cloud providers I can trust completely, which will not share my data with anyone else without my say-so.
But in the rush for companies to monetize data by turning it into information about particular demographics, I think we’re missing the point. The greatest value of that data is to me, not to other people. Say, if it is health data, which suggests I should change my lifestyle to increase my longevity.
There is also data I want to analyze.But to do that, I have to be able to integrate all that information. And I want to be able to do that on my own computer, or at least I’d like to have a choice of where that happens.
This all relates to your recent work at MIT.
What we’re doing in the lab at MIT is building apps that store data on a generic piece of storage. Instead of each app having its own API, you have a generic API and the data is stored by the apps into a generic piece of storage. I can tell the app where I want it to store the data. I have control over it and I make my choices about where it gets backed up and who it gets shared with. And crucially, I get the extra power, which I think people underestimate, of really being able to do lots of things with their data.
So, what does one of those apps look like?
One application area where this is done quite well is calendaring. CalDav is a standard protocol that lets calendars communicate. With iCalendar on the Mac or your iPhone, you can bring in calendars from different places. I’ve got one stored in the cloud which is my personal one, which I share with a number of people in the family. I’ve got several provided by W3.org which are shared with different groups. And I’ve got my work calendar, which I share with my team and family. I also bring in calendars for bands that I like and so on, so when I am deciding what to do, I can look at all these things.
It’s like that for calendars, but it’s not like that for maps. If I want to look at a run I’ve just been on and you’ve just been on a run and I want to compare it and see if we’ve been on the same run, maybe with all the runs that people in my running club have been on, and I want to choose that I look at it in Open Street Map, we can’t do that. Instead, each map is a different web site and my run is in a different place.
But basically, in the future, everything will be social, so everything will be like a calendar: Data-driven, compiling data from different social groups, which will have different access control.
ᴙғ: One of the things I’ve been thinking a lot about is how mobile apps, which run around the web, but are not fully part of it, are succeeding because they are just so convenient, even if they don’t have all the features and openness we’re talking about. Should one of the primary goals of the Web Consortium or your new projects be to increase the raw convenience of the web and make it as slick as these mobile apps?
Yeah. We’ve put a lot of energy into this. There is the idea of a web page being just like an app. It has access to all the things on your device, the camera and so on. It’s got access to really interesting data. It’s figuring out how to do the control access to allow you to decide which apps you’re going to share it with.
One of the ideas that some people have been talking about is a beneficent app. It is like benevolent, but it’s the word you use when you’re trying to do a human-factors experiment. The ethical committee at MIT says, “Are you trying to cure the disease or give someone the disease? If what you’re trying to do is help people, then let’s do this experiment.”
The idea of basically being beneficent is one that’s really important. It used to be that all the apps that I bought for my laptop, they were there to service me. They don’t advertise. They don’t sell all my data. At one point, Quicken, the money management app, said, “Do you want to buy some insurance?” And I said, “I paid for you to be my agent, don’t you dare advertise to me!”
If you are a beneficent app, the developer must say, “I will not steal any information. If I do aggregate information, it will be only to provide useful feedback to the user. I will always do what I would have wanted as the user. There will be no monetizing of data, no advertising. There will be no trying to twist the user to move in a particular direction politically or religiously.”
That used to be the default. So many apps were built like that. And now people are starting to assume that nothing will be like that. So, if we could have that branding of beneficence, that would make life a lot easier because you could tell your computer, "These are my beneficent apps, so go ahead, and give them access to financial details, and so on."
ᴙғ: It gets at this fascinating concept that we have really different relationships with all the the apps co-existing on our phones. Even if we use them all together, there are really different standards of trust at play.
A student here did an analysis. She looked at the Google Play store, looking at which apps ask for what personal information. And of the apps where there was a free and paid version, by far the most common thing that the free version did that the paid version didn’t was your calendar. If you think about it, if you want access to one thing that will allow you to completely profile a person, the calendar has an incredible amount of personal information about who you are seeing, medical things, and so on. That’s the sad state of affairs. Paying for app development by sneaking a look at my calendar is no way to run things.
ᴙғ: You’ve talked a lot in the past about the web growing up as a non-national entity. But it seems to me that the past few years have seen a resurgence of techno-nationalism, officially in China and Spain and off-the-books in the US in terms of our spying infrastructure. Does this surprise you?
The Internet is inherently non-national. Nations are not part of the way the Internet is designed or the web is designed. When you click on a link, you really have no idea of national borders. Where stuff is stored is independent of what domain name it has on it or the nationality of the person responsible for it, and so on.
The web has been a wonderful renaissance in part because of the lack of national borders.
The whole system is governed by code and by law. Both are important. And the idea that either should be made without the other is not tenable. To have people trying to make laws, about, say, cookies—there could have been a really interesting law that was about third-party cookies, but not first-party cookies, but that needed to be done by technologists working with regulators.
Or take the Digital Millennium Copyright Act, the way it puts people in jail for tens of years, like Aaron Schwartz faced, in a completely inappropriate way. Really bad things happen when laws are done without understanding the technology. We can design new technological systems, and when people design technology, they need to think about what effect it may have on society. And nowadays, I am encouraging people to think about what sort of laws will be necessary.
We’re talking at the moment of the Magna Carta for the Internet, the Magna Carta for the Web, and we’re talking about doing that as an international thing. We’ve had a lot of discussion around the celebration of the 25 years of the web about it.
And the Magna Carta is full of general things which are becoming human rights like the ability to connect with whoever you want across the Internet. The ability to do a web search, because if you don’t have it, you are so disenfranchised. Things like that which you wouldn’t have thought of as human rights.
But in every jurisdiction, you need to have separate teams working on each country evaluating how well they’re doing in terms of the openness of the web.