more on distributing computing

Developing good social networking technology requires thinking about distributing computing between humans and computers. Way back in 2002, a human-computer interface designer discussed some problems with the then-trendy idea of context-aware computing:

I suggest rather than trying to take humans out of the control loop, we keep them in the loop. Computational systems are good at gathering and aggregating data; humans are good at recognizing contexts and determining what is appropriate. Let each do what each is good at.
[Erickson (2002) p. 103]

Recognizing and respecting comparative advantage between humans and computers is also a good design principle for social networking technology.

Social networking technology that depends on a computer having better (human) social intelligence than a human challenges the designs of both. Consider a social networking application for a mobile phone called a Jerk-O-Meter. It measures the user’s voice activity and voice stress. Using these data, the application evaluates the user’s communicative performance and delivers these messages:

“Stop being a Jerk!”
“You could do better”
“Now we’re getting somewhere”
“Wow you’re a smooth talker”
[Madan & Pentland (2006), p. 6]

Another application, called Wingman3G, measures speaking time, voice rate, and vocal stress. It evaluates this data using a model of successful dating communication and produces real-time messages such as:

“Maybe you could speak a little slower?”
“You’re getting there, maybe you could relax a little?”
[Madan & Pentland (2006), p. 7]

Human brains evolved under selection for social intelligence. Digital computers did not. Human social intelligence can easily encompass that of computers and reduce their social value to the social value of recognized manners and conventions.

Compared to humans, digital systems are relatively good at routine collecting, processing, and distributing information. Information such as on-line/off-line status, communication initiation, communication addressing, communication duration, as well as word rate and stress indicators, might be valuable to humans using social networking technologies. An interesting recent paper discusses the design of shared visualizations of such information (“social proxies”). It offers six claims for good design of social proxies:

1) “Everyone sees the same thing; no user customization”
2) “Portray actions, not interpretation”
3) “Social proxies should allow deception”
4) “Support micro/macro readings”
5) “Ambiguity is useful: suggest rather than inform”
6) “Use a third-person point of view”
[see Erickson (2006), pp. 13-4, which describes these claims in more detail]

Claims 1) and 6) suggest designing social proxies to be like objects in our one, real world. Claims 2)-5) point to the comparative advantage of human social intelligence in human social interactions.

* * *


Erickson, Thomas (2002), “Some Problems with the Notion of Context-Aware Computing,” Communications of the ACM, v. 45 no. 2 (Feb.) pp 102-4.

Erickson, Thomas (2006), “‘Social’ Systems: Designing Digital Systems that Support Social Intelligence” (pdf file).

Madan, Anmol and Alex “Sandy” Pentland (2006), VibeFones: Socially Aware Mobile Phones (pdf file).

Leave a Reply

Your email address will not be published.

Current month ye@r day *