Update 2007: Prior to the talk, I hadn’t heard of Adam Greenfield. But in time, I learned that he is one of the leading experts on this subject. Even after SXSW2007 I haven’t attended a more stimulating panel than this. I’ve cleaned up this post so it’s a bit more coherent. My own insights are encased in [ brackets ].

What is ubiquitous computing?

Greenfield starts by defining the subject. He gives an example of devices found in three main areas:

1. Ubiquitous devices on the body. Example: BodyMedia SenseWear (this device is strapped around the arm, much like an iPod nano armband).
2. Ubiquitous devices in the room. Example: Sensacell. It looks like a grid on the table/floor.
3. Lastly, ubiquitous devices on the street. Example: A pan-optic surveillance system (for prisons). Prisoners in cells are not aware that they’re being watched.

Everyware is different. So it’s incredibly difficult to design and plan for. People using everyware are not (knowingly) engaged in a technical interaction at all. There are many devices, system and services impinging on each user. Their positions (locations, orientations) matter. There may be social or political meanings associated with where they are. Greenfield states that this system is very complex.

Greenfield gives the example of a man he saw who was unable to find film in a supermarket. —> What about using RFID tags for this?

The challenge with switching to ubiquitous computing

When you interact with a personal computer, it knows that the commands you’re giving, are meant for it. However it’s not the same with ubiquitous computing. What if you’re making a transaction and you suddenly swipe at a fly?(!)

He gives the example of New Songdo, 2004 - a city built with RFID. It has a systems in the floorboards of housing for the elderly to detect if someone’s fallen down. It will then open the front door and call for assistance.

Look also at Mastercard paypass. It has a chip embedded in it. [Hey that’s like the Esso-Mobile SpeedPass I use back in Singapore to pay my bills at the petrol station.]

At this point, Greenfield says something like, “Frankly I have no interest in Web 2.0. XHTML, CSS - how much longer do you want to hang around, until you sink your teeth into a new challenge?” i.e.Everyware!

The Misery of User Experience

Greenfield gives some examples of this. Showerheads that need adjusting in the morning. Toilets that ‘crash’ [Hmm he means it needs plumbing? ;-)] Basically, we have a long way to go.

Also, information technology requires details that are explicit, whereas in real life it’s not always like that. You can do wireframing for websites. There’s no equivalent for everyware.

Greenfields mentions an article/query he posted in Boxesandarrows.com on ubiquitous computing, which drew comments from both ends of the spectrum. [This could be the one he’s talking about: this article.]

Principle 1 - Default to harmlessness.

Systems do fail. When they fail, they should default to a mode that ensures users’ safety - physical, psychic and financial. He calls it ‘graceful degradation’ [Hey, that sounds like CSS! :-P ]

Principle 2 - Be self-disclosing.

Ubiquitous systems must contain provisions for immediate and transparent querying of their ownership, use, capabilities, etc. Let’s go back to the example of New Songdo when an elderly patient falls on the floor. Let’s say that the patient didn’t want people to know she had fallen (she wants some privacy). Why should the system open the door to strangers?

Seamlessness must be optional mode of presentation, not a mandatory or inescapable one.

Principle 3 - Allow people to save face

Ubiquitous systems must not unnecessarily embarass, humiliate or shame their users. Eg Flickr.com. Your friend has over a hundred friends and two contacts, and you’re one of them. You wonder why you aren’t good enough to be listed as a ‘friend’. Such systems don’t account for these uncomfortable situations.

Greenfield cites a true story of a man who wandered to the wrong part of a building, was detected via RFID and his error was announced over the PA system, along with his full name. Greenfield asks, “Why humiliate when you only need to correct?” [This principle would be VERY important in Asia!]

Principle 4 - Be conservative of time

Greenfield’s example: “If I want a cup of tea and I put the kettle on the stove, I do not want the kettle to ask me, ‘Do you really want to boil this water?” [i.e. Just do it?]

Principle 5 - Be deniable

Ubiquitous systems must offer users the ability to opt out, always and at any point. People shouldn’t be relegated to the back of the bus, just because they don’t want to use the system. [I’m thinking about how we in Singapore switched from paying coins on the bus to using a cashless card system, and how those who continued paying with coins had to pay a bit more than the rest.]

Questions

Question 1: How do you redesign your infrastructure (e.g. Wifi, broadband) to adapt to ubiquitous systems?

Greenfield imagines that mesh and ad-hoc networks will play an important role in future. He’s not an expert on this part.

Question 2: What about a bottom-up approach? ie, letting the people actually using the system, have a say in designing it.

Greenfield says we need to strike a balance between bottom-up and top-down evaluations. He’s always been inspired by the Web Standards project in this aspect. They found a business logic to promote their cause. [i.e. we should be in touch with our markets.]

Question 3: Will there be a rise in technological resistance? Eg. home-made RFID blockers.

Greenfield says yes! People who know how to do it, will take action in their own hands.

My rating: 4.5/5. This was the one panel where I learnt the most new things and actually felt like starting something new.