TCS Daily

Social Software

By Arnold Kling - April 21, 2003 12:00 AM

In 2003, it feels as if someone hit the "pause" button on the innovation machine that drove the economy for the previous twenty years. The existing infrastructure is sufficient to support the applications of word processing, spreadsheets, email, and web browsing. Where is the next generation of "killer applications" that will drive mainstream adoption of technologies that are tantalizingly close to realization, such as wireless Internet access, pervasive computing, and radio on a chip?

A reasonable guess is that the next wave of compelling applications will be in the realm of what Clay Shirky and others call social software. There are many challenges in shifting from software design that focuses on individual use to software design that focuses on group use. For example, Shirky writes,

"If a group has a goal, how can we understand the way the software supports that goal? This is a complicated question, not least because the conditions that foster good group work, such as clear decision- making process, may well upset some of the individual participants. Most of our methods for soliciting user feedback assume, usually implicitly, that the individual's reaction to the software is the critical factor. This tilts software and interface design towards single-user assumptions, even when the software's most important user is a group."

Perhaps the best way to explain the issues surrounding social software development is to describe the types of problems that such software might solve. These might be called the matching problem, the issue-resolution problem, and the classroom-management problem.

The Matching Problem

The matching problem shows up in a variety of contexts. In the labor market, the problem is to match hiring managers with qualified workers. When you need something fixed, the problem is to match your dented fender or stopped-up toilet with the appropriate repair person. When you want to go out on a date, the problem is to match you with someone who is compatible.

Matching requires two basic steps. One is to gather a list of possible matches. Other things equal, it is better to have the largest possible list, in order to reduce the possibility of failing to find a match. The challenge with this step is to have sufficient reach to include all plausible matches.

The second step is to filter the possible matches, so that the most likely matches come to the surface. The challenge with this step is to come up with effective filtering algorithms. For example, many job search sites match jobs and candidates on the basis of key words. That is not a very intelligent approach. Personal recommendations often work best in solving a matching problem. I would like to find matches based on the recommendations of friends who know my tastes or strangers who happen to share my tastes.

The challenge for social software is to manage a large social network for obtaining potential matches while implementing effective filtering algorithms for providing helpful personal recommendations. Many solutions exist that are partial and less-than-perfectly satisfying. There is room for dramatic improvements in usability and effectiveness.

The Issue-Resolution Problem

The recent wars in Afghanistan and Iraq saw a speed-up in the process of resolving the issue of hitting an enemy target. As Fred Kaplan reported,

"A new generation of unmanned Predator drones flew over the battlefield, scanning the terrain with digital cameras and instantly transmitting the imagery back to command headquarters. Commanders would view the imagery, look for targets, and order pilots in the area to attack the targets. The pilots would punch the target's coordinates into the smart bomb's GPS receiver. The bomb would home in on the target. Total time elapsed: about 20 minutes. By comparison, in Desert Storm, the process of spotting a new target, assigning a weapon to hit it, then hitting it, took three days."

Corporate America could use a similar speed-up in completing projects that cut across functional lines. Examples of such projects might be developing a new web site or launching a new product or service.

The problem posed by these sorts of projects is combining the work of a full-time project team with the knowledge of line managers who have other responsibilities. The project team cannot resolve issues on its own without making serious mistakes. However, the team cannot obtain the expert guidance that they need from multiple line managers without slowing down the project.

The challenge for social software is to keep important managers involved in projects without taking them away from their line responsibilities. Like combat air support, they need to be available when needed. This means not only good communications systems but also systems for managing priorities.

The Classroom-Management Problem

When I teach, I find that students are multi-tasking. The pace at which I teach is bound to be too slow for some and too fast for others. Some of them may be working on other subjects, and some of them may be daydreaming or conversing. In the corporate world, you can see the same thing happening at business meetings, where even the colorful cartoons of a PowerPoint presentation often fail to hold the attention of the majority of the audience.

How can we improve the delivery of material from teachers to students? Teaching equals feedback, and the best feedback tends to be individual feedback. Teachers get feedback when they see how students do on tests. But we are always looking for real-time feedback, whether it comes from students raising their hands to answer questions that we pose or from puzzled looks on their faces.

The challenge for social software in classroom management is to enable the teacher to be aware of what the students have grasped and what they still do not understand. The teacher then may need to deal with different subsets of students in different ways.

For many of the challenges of social software, some components appear to exist. We have some tools to filter recommendations. We have some tools to prioritize messages. We have some tools to provide and monitor feedback. However, we do not yet have the components sufficiently tuned and assembled into applications that solve the larger human problems. My guess is that at least one important innovation of the next decade will be a system that solves one or more of the problems that I have sketched out--the matching problem, the issue-resolution problem, and the classroom-management problem.

TCS Daily Archives