XTech 2007: The Ubiquitous Web

Welcome to XTech in Paris, a conference that looks just over the horizon into the near future or, in many cases, the advanced present of the internet. This is not your parents’ internet. It is an internet freed from the not only the desktop, but the computer entirely.

The schedule for the conference is here. If there is anything you’d like to know or anyone you’d like to put a question to, leave a comment, and I’ll do my best to get you some answers. WiFi is a little scarce here, oddly, so I’ll be online as much as possible but not as much as I’d like. Everywhere internet, design, location-based services, web app development is just a taster of the topics covered.

Software becomes ‘Everyware’

Adam Greenfield kicked off the second day with a keynote that looked at the broader implications of the increasing reach of networked technology in our lives. (I missed the first day, but you can find a lot of posts by searching XTech in Technorati.) He said that he comes from the confrontation of human beings with technology. He looks at how people use technology, and he says that he feels their pain.

What is ubiquitous or pervasive computing? He said that very little of what he was going to talk about will deal with the web. He wanted to talk about what he calls Everyware (Get it? Software goes everywhere.) Back in 1990, Mark Weiser thought about embedded, wireless computing that went far beyond the desktop computer, beyond a GUI. It automates and digitises all kinds of ‘unheroic tasks of everyday life’.

He gave examples such as a digital doorlock that is part of everyday life in Korea. You can use bluetooth, a biometric scanner or even a plain old key to open your door. Technology is ‘going to the body’, and he gave the example of the Nike+ iPod, where not only would it gather your exercise information but allow you to share that with others. The world becomes your exercise partner.

There is an emergent ‘internet of things’. All of these services communicate not just with human beings but also with each other.

A class of systems tends to colonise everyday life. We’re in France, so Adam gave a reference to Michel Foucault’s idea of the Panopticon. The prisoners never knew whether they were being surveilled so they had to assume that they were always being surveilled.

How does this impact us now? He gave the example of the Kinko, a networked toilet. It would analyse your bodily wastes and transmit information to your doctor. Surveillance becomes not just the watchful gaze but a product of the technological systems deployed in everyday life that gather a lot of data.

Is this all science fiction? No, he believes that this is a present, real-world concern. Most of the time he has had to illustrate his talk with prototypes, but the systems are now becoming real-world, commercially available products.

For example, he says that the new internet protocol, the 128-bit address space IPv6, provides some 6.5 to 10 to 23 power addresses for every square meter of the earth. You can give an internet address to almost every object in the world. Seem ludicrous? He gave the example of proliphix.com, a service for networked thermostats.

He’s worried that most people see ubiquitous computing as unproblematic. What do you have to consider? Inadvertency. They didn’t mean to engage this system. Now, what happens when people don’t know about the system or are unwilling to engage with that system? And with all of these interlocking systems, there may be a lot of unintended consequences. One system triggers another and causes a cascade of unintended and unwanted results in systems around it.

It’s time to take everyware seriously, he says. He laid down a few principles that he believes need to be considered as these systems are developed (someone in the audience echoed what I thought that his principles mirrored Asmimov’s laws of robotics).

Principle 1: Default to harmlessness. Ensure users’ physical, psychic and financial safety, but realise that means different things to different cultures.

Principle 2: Be self-disclosing. Seamlessness must be optional, and it has to be clear who owns them, what they use and what they do.

Principle 3: Be conservative of face. They must not embarrass, humiliate or shame their users. We have to build systems that operate a high degree of precision but then ‘fog’ those results in imprecision.

Principle 4: Be conservative of time. Don’t make life more complicated than it already is.

Principle 5: Be deniable. People have to be able to say ‘no’.

Open API to counter climate change

Gavin Starks with d:gen networks spoke about climate change and quoted a statistic from World Changing. We ship 2.4 trillion kg of cargo from port to port per year. It’s like shipping every human being six times. “Isn’t that insane?”, he asks.

Temperatures are expected to rise between 1.8 to 4 degrees Celsius, according to the Climate Group. The last time this happened 125,000 years ago, the sea was 4 to 6 metres higher. We are looking at hundreds of millions of people under threat. The Stern Review on the Economics of Climate Change said that greenhouse gas emissions need to be 75% lower by 2050.

He argued that 75% of that reduction could come from current technologies. We could get a quarter of the way through conservation and increased efficiency. Wouldn’t it be great if we could shut down some power stations? Two years ago, he was asked: “How would you get a billion people to reduce their carbon emissions by one tonne per year?”

I can’t just click somewhere to absolve my life if I want to reduce my energy consumption by 85%. We need more than a campaign. We need a movement. We are missing a development framework. What would we build? What should we do?

There needs to be trust in the information. To scale, we need to share tools. We need to give away the tools and share things. We don’t need to re-invent the wheel, build another carbon calculator. We need an available engine, that we can trust and share and remembers what I’ve done. It needs to be extensible and open.

In December of last year, he got a call from the UK government and got all the data and calculations from the government. They have their own scientists who add data, and they source every single item of data. They have created a ‘profiling engine’. Anyone can run a campaign and use anonymous keys to share stats. They are giving it away for free via GPL.

He sees toolkits for schools, builders. Use RFID versions to ‘track the trackers’. He sees league tables for the good and the bad. Data sets over districts, regions and campaigns to create competitions to cut carbon emissions. He wants to integrate it with Pledgebank. How do you integrate this with Google, Flickr, Dopplr, Twitter, Make or NagMe?

They have a generic engine with generic anonymous keys that allow the sharing of data.

Let us be very clear that it’s not the planet we’re saving, but its species and most notably our own.

It is launching today. The site it launching today. Go to http://www.dgen.net/amee

AMEE=Avoiding Mass Extinction Engine

There are database issues and database rights issues. We have some potentially big privacy issues. We have done our best to keep it anonymous, but there is still a need to protect this information.

Ian Forrester, with BBC Backstage, asked what public broadcasters like the BBC can do, and Gavin said, “You already are.” He would not go into details.

Matt Biddulph is building Dopplr, a traveller site that allows people to show where they are travelling. It helps people let their friends know if they will be in the same city at the same time.

“It is not a competition,” Matt said. He wants to know how the site might work with this service. The site is in a private beta at the moment. (I use it, but I think I’ve given away all of my invitations.) Dopplr users have booked 9 million miles of air travel, Matt added.

Jabber: Social Software for Robots

Blaine Cook, Obvious Corp

Kellan Elliot-McCrea, Flickr/Yahoo!

Flickr and Twitter. We say we put the point in Web 2.0, they say.

Social software is people asking computers to talk to people. Your actions are aggregated in one place.

It took me a while to get my head around what they were talking about, mostly because I use web services. I can’t code. After a while, I figured it out. They were talking about using messaging protocols, in this case the Jabber messaging service, for computers to talk to each other rather than people chatting online. It gets around some of the problems of pushing around real-time information between computers without the computers idling and waiting for information. The ‘Are we there yet?’ problem.

There are big clouds of data floating around, and the current models are breaking down. “It makes computers cry”, Kellan said, adding, “We say at Flickr that we make computers cry everyday.”

Let’s consider a new model for web services. Message passing through asynchronous communications. The computer asks: Let me know when we’re there. When it is, the server says, “We’re there”. Social software for robots is message passing.

Where it works? It’s not actually useful for everything, but it’s good for real time and wire level data with no database interaction. It doesn’t work so much with historical and static content or interactive searches.

What we’re thinking about is a real-time Craig’s List. In San Francisco, it’s really important if you want to rent an apartment. If you could see that listing when you’re out on the street, you can get the apartment. But if you poll the IP every three seconds, then you’ll be blocked under current models.

What happens when you’re fishing for information? Tell a little bot that you want to know when a new apartment is listed on Craig’s List. You put a net out, and you’ll get notified when you catch something. It’s important when you’re waiting for something to happen as opposed to looking for historical information.

How do you build it? Jabber. It’s usually considered a messaging protocol, but it’s a standard. You can do internet scale message passing. It’s just XML. You can do this securely and verify the identity of the sender. No spoofing!

This brings up a whole new set of services. Private, secure services.

“It’s hard to talk about how cool this?” We can have the server asking the client what’s new.

If a server is processing a 100MB wav file, the client gets in the way of the processing by asking “Are we there yet?” Is the processing done? Instead, they can use Jabber to pass the message of when the task is completed instead the client pestering the server needlessly.

Jabber is asynchronous, real-time, extensible, secure, delegated. It has callbacks. It’s standardised. It has presence. It is decentralised but not P2P. (Sorry, I’m going to leave that one hanging there for the code-literate amongst you.)

The http protocol, one of the most common protocols on the web, just doesn’t work for everything. With real-time information, the number of data calls quickly becomes problematic. This is a novel way to push information around which

Jaiku: Rich Presence.

Ralph Meijer: Jaiku

We used to use mobile phones just to call people. Later on, you had contact lists to make it faster to call them. But how do you know if they are there? If they are available? Maybe the person you’re trying to call is in the theatre, busy or driving.

But now you can have pervasive net connectivity. You can find out where a person is. You can have many-to-many communications. They have written Jaiku for S60 phones, the OS for many Nokia phones. You can see your contacts and what the person was doing recently. Your contacts can know what you are doing and if you are busy.

They use cell tower information and bluetooth neighbourhood information to give a sense of presence, some location information. They have worked to integrate calendar data. On some of the newer mobile phones, you can show your friends your photos to allow them to see what you are seeing. They are working to balance battery use and the need for frequent if not persistent connection to the mobile data network.

Jaiku also is on the web. You can send messages. You can also pull in web feeds from your blog, Flickr, last.fm or Twitter. People selectively subscribe to your feeds to keep a balance between knowing what their friends are doing and being overwhelmed with information. You can add comments to other people’s blog posts or Flickr pictures.

Jaiku pulls together a lot of the social applications through RSS and aggregates this information in one place as well as making this ‘lifestream’ mobile. They also are working on an SMS service. They have created ‘channels’ or groups. They recently did this for the Eurovision song contest.

They use most modern web technologies like JSON, RSS and Atom, different ways to distribute all kinds of data and information such as blog posts, a listing of your latest Flickr photos or your podcasts.

He is asked about something between location and presence so letting someone know you are home without specifying where home is. Ralph says that the XMPP protocol, which is used quite a bit in Jaiku, allows you to specify how much information that you’d like to publish. You don’t have to publish exact GPS coordinates. You can be as precise or fuzzy as you’d like.

On Jaiku, you can also set your messages to be public or private to determine how widely the information is available, to everyone or just your contacts.

Why did they include RSS? We think your personal feeds are part of the complete picture of what you are up to.

What about the data prices? Some of these services can cost a lot of money with all of the data being passed.

We are talking with various telco providers to arrange something in that area. SMS aren’t very expensive. In the Netherlands and Finland, you see smaller providers providing flat fee data plans. We are counting on always-on data plans.

Technorati Tags: , , ,