Thursday, 10 November 2011

Context computing

A good session this morning on context computing.
Context is based on a set of principles and technologies designed to make services more usable, relevant and fun. Using more information sources, more social information leads to  hyper personalisation of services.  It's the integration of mobile, social, digital and physical worlds.
Theres a lot of it already about. Face recognition and emotional detection already being used on vending machines in Japan so they can make judgements about what you might want. 
Smart products exist, eg glasses with electronics in so know when full or empty and can order your next drink.

Consumers are motivated by emotions. Detection of emotions will be increasingly important in next few years.

By 2014 40% of smart phone users will opt in to context services that track you. Will trade off some privacy for better services.

Last decade of Internet dominated by search, a pull technology. Next decade will be dominated by proactive push technologies.  Will be personalised and involve social information and be  multi channel.
Information will be key, and will be collected in  4 key areas:
Intent. What does user want or want to do?
Environment. The current state of user. Where are they, what are they doing. Social dimension, community
 Identity, validating who you are, reputation information.

 Will be very complex, and one of the biggest challenges will be that all this information comes from different places. Currently no standards. Will require sophisticated information models.

New technologies being developed such as Emotion ML, a mark up language to tag events with emotional context.

Mobile consumer application platforms becoming increasingly important as most of the information necessary for contextual computing  will come from mobile devices.  Many different architectures exist, Apple and Google are taking an early lead, but lots of others coming up eg Appcelerator and Kony. Gartner have a  magic quadrant for MCAPs which is worth a look.

Ensemble interaction being developed, ie  interactions that cross more than  one device. Eg TV and mobiles. So, you could be watching TV, like the look of some clothes someone is wearing,  point mobile phone camera at screen, find what they're wearing and buy it.  This is already possible.  A  NFC (near field communication) enabled phone will pick up information from smart posters. 
Smart posters exist  in Japan with web cam in it, looks at you and decides about what sort of advert to show you and enables you to pick up a discount voucher on your phone. 

Starhub have developed  smart changing rooms which detect the RFID tag in clothes you're trying on and chooses the music to play. Hmm, not sure about that one!

Social media very important to context. Can deduce influencers.already developing " Pay with a tweet" for music downloads, Ie pay for a music track by tweeting about it. 

Managing risk, the dark side of context aware computing.
Gartner predicts that Google, Microsoft, Nokia and Apple will continuously track daily journeys and digital habits for 10% of the worl'ds population by 2015.
Some people will find this creepy and won't want it. Some will see that it gives them better services and see the advantages. 
Privacy will be important. Need simple, transparent privacy policies, and easy ways to opt out, and to correct deductions about behaviour.

Good session to finish the conference on, and some interesting things on the horizon. 

No comments: