The Age of the Mobile Mash-Up

By Lars Erik Holmquist of the Mobile Life Centre, Kista, Sweden

The rate of innovation in mobile services is just about to take a quantum leap. We are going from a divergent and messy ecosystem, where every new concept has to be made into a specialized ”app” that works only on a small sub-set of mobile handsets (even the mighty iPhone only has around 3% of the global mobile phone market), to an environment much more like the web. Today, new services can easily be composed out of existing components and run on a common platform – the browser. We are entering an age where the creation of a new mobile service – taking advantage of such features as the user’s location, social network, personal data, and even phone-specific functions such as the camera and accelerometer – can be mashed-up and put on-line just as easily as Web 2.0 services have been for several years already.

At the Mobile Life research centre in Kista, Sweden, partners from academia and industry are working together to imagine this future of abundant mobile services. Fortunately, we are not working in the dark – we can build on a foundation of several decades of research. Some 20 years ago, Mark Weiser, a research scientist at Xerox PARC in Palo Alto, had a vision of the future: he called it ubiquitous computing. He imagined that dozens, even hundred of small computers would be available everywhere, and seamlessly support us in our everyday tasks. Unlike the personal computers at the time, these devices would be un-tethered, user-friendly, aware of their surroundings, and conducive to communication and collaboration in the real world rather than through a screen. To explore this vision, he and his team built a number of computing devices in different sizes – they called them Tabs, Pads, and Boards. Each was connected to a wireless network and aware of its location and other factors in the environment, the so-called context.

Today, this vision is reality. The mobile phone, in particular smart phones such as the iPhone, Blackberry and Nexus One, are doing what the Tab did and much more. In practice, Pads have also been with us for a long time in the form of highly portable laptops and netbooks, although it took Apple to come out with a keyboard-less laptop device that actually worked (and they named it with more than a nod to Weiser). Interactive Boards have been less successful, and so far we are mostly seeing wall-size displays in the form of PowerPoint presentations and non-responsive information screens.

But while the devices are more or less the same, today we have something that the researchers at PARC did not: Infrastructure. In 1990, to build an always-on, network-connected, context-aware device, you would have to start from scratch – not just by putting together the device itself, but also with laboriously setting up radio and infrared base stations in every single room. Today, you buy a phone or laptop and hop on the 3G network or find the nearest Wi-Fi hotspot. That today’s devices are hundreds of times more powerful when it comes to processing power and memory is really not so important in terms of the software you can now create for them – many of the application concepts that are now popular on phones and pads have been around in the research world for decades. What is different, and what is truly a gamechanger, is that every device is connected, not just to other devices, but to rapidly growing resources of software resources, social networks, and data.

So what does this mean? Well, first of all that every ubiquitous computing service dreamed up by researchers in the last 20 years can not only be implemented on a consumer device, but also reach critical mass very quickly. For instance, a decade ago, I thought about the problem of sharing data between devices that had no obvious user interface for connection, but shared the same radio network. My idea was to create a shared context for the devices by holding them together and shaking them. A type of sensor called an accelerometer (now available in most phones) could pick up the direction and speed of the movement. Since the chances were slim that any to devices in the same place would be having exactly the same motion characteristics, unless they were indeed held together, they could securely open a connection between them and send over data, for instance a business card. Today, an app called Bump does exactly this but cleverly adapted to a new infrastructure. Rather than relying on the devices being in the vicinity of each other through radio connection, it uses the phone’s location information to determine if two devices that have the same motion characteristics are also in the same place – and thus being bumped or shaken together. We published the sharing technique in 2001 (as Smart-Its Friends), but while we only ever built a few devices to demonstrate the concept, Bump is available through various app stores and numbers millions of users.

Furthermore, with the accelerating capabilities of HTML 5 and mobile browsers, soon we will not even have to go through the process of installing apps on the phone – everything will be on the web. Let me take one other example. Over ten years ago, I was thinking about how one of the PARC projects – the Active Badge – could be made truly mobile. This was a device that was worn like a badge, but it contained some clever electronics, including a transmitter of infrared (invisible) light. Corresponding infrared receivers were located in every room of the office and wired to the local network, so that whenever somebody entered a room, the sensor would pick it up and transmit the position of the badge wearer. Using this information it was possible to deduce not only where a person was but also high-level information like for instance if a meeting was going on (several people at the same time in one location), to log a person in automatically to a workstation when he or she entered a room, and even route fixed-line phone calls to the phone nearest the recipient rather than their personal phone.

The Active Badge performed a useful service, in that it kept track of the location of every badge wearer in the office. But it not only required the custom built ”smart” badges, it also required a very laborious infrastructure, with an infrared receiver located in every room of the building. What if it could be re-fitted so that it did not require a permanent installation of beacons and networks? Our solution was the Hummingbird, a device which tells you which of your other friends are in the vicinity – provided, of course, that they also have the device. In stead of relying of fixed receivers, it did this by listening to the radio transmissions of other Hummingbird devices, and since each was tied to a specific person, it was easy to deduce that the wearer was close (as in a couple of hundred meters) if the signal was picked up. A small display on the device showed the name of nearby users and a soft ”hum” was heard whenever someone was close. The downside was that the wearer’s absolute location was never known, but the upside was that it provided the really useful bit of information – who is near me? – without the need to install any additional hardware.

Location-aware ”friend finder” technology like the Hummingbird is now available in hundreds of services, including major ones like Foursquare and Gowalla, and some of them have a millions users or more. But they still require the installation of an app, and this will always be a hurdle for reaching users. Instead of working for a specific platform, through the location feature of HTML 5, it is now possible to take advantage of location features in the browser. Thus, a location-based service can now often be implemented in a matter of days on the server side, and be available instantaneously in all phones with compliant browsers – which soon will be pretty much all of them. For instance, my lab just implemented a mobile web version of the Hummingbird, whereby visiting a web page you get a list of the users close to you in order of distance. No installation is required, and it works just fine on mobiles as well as iPads and PCs (we decided to skip the ”hum” this time!) While the native app explosion has already been phenomenal, I believe this takes mobile services much closer to the innovation pace and critical mass potential that has been prevalent on the web for the last decade.

But in my view this new infrastructure is not just radio waves and wires – there is also the abundance of software components and social networks. Years before the iPhone and Zune, my group worked on social mobile music applications. We came up with Push!Music, where songs would move between devices all by themselves using ad-hoc radio connections, to find the environment where they would be most appreciated. Say you listen to a lot of Johnny Cash and country, but happened to download some Metallica by mistake. One day you share a bus ride or sit in a café next to a heavy metal fan – and those unloved Metallica songs may decide to jump ship from your player to one where they have a chance of being more appreciated! To test this concept, we had to buy expensive handheld computers with Wi-Fi and loan them to test subjects, who used them in lieu of their ordinary MP3 players. Because of this, we could only ever have 10 or so users at a time, and achieving ”critical mass” was never possible.

Today, of course, we could have done the same as an app and reached many more users. But the opportunities even richer if we consider social networks and existing music services as part of the infrastructure. Using a little bit of clever coding it is possible to glue them together to create entirely new concepts in no time. To prove the point, two of my colleagues returned from an extended weekend of coding with SpotiSquare – a mash-up of Foursquare and Spotify. It allows you to do a very simple but powerful connection: tie a Spotify playlist to a Foursquare location. Thus any café, office or venue can have their own dedicated playlist, available to listen to and modify when you check in. And since it is again a web page, it is compatible with almost any device. This is not just a re-hash of something researchers could have thought of in the last century – it is something new, because it builds on an ecology of practices and services that never existed before.

This is where I believe the most exciting opportunities lie right now in mobile applications: picking and mixing from hundreds of mobile web services and social networks, moving the grunt work from the handsets to the cloud, and adding a dash of location and context-awareness. Of course, it does not mean you will also automatically have a viable a business plan – as we know, a great technical idea is only a small part of what it takes to build a business – but it means that the step from thought to users is shorter than ever. We are finally entering the age of the mobile mash-up, and it will all be happening at even greater speed than the Web 2.0 revolution. 20 years after Weiser’s vision we are ready to take the next step, beyond ubiquitous computing, and it is up to researchers and entrepreneurs to invent the future once again.