Archive

Archive for September, 2017

Computing Challenges We Can’t Solve with Research: Simple Ideas and Platform Lock

September 22, 2017 3 comments

Critical problems for computing and society are increasingly economic. It’s not that we don’t know how to fix them—it’s that a purely market economy model to fund software development doesn’t support some simple things that would make the world a better place.

My students and I research and design collaborative computing systems. We start by studying existing groups and trying to understand their needs. For example, Kurt Luther studied animators trying to work on collaborative projects with teams of fifty or more artists. Based on what he learned, he built a project management tool for creative collaboration that tried to balance decentralized and centralized creative decision making. Jill Dimond worked with the nonprofit Hollaback, which is trying to end street harassment, and helped redesign their website with a federated structure that helped them to rapidly spread around the world. In these cases, as researchers were able to understand the problem, and innovate to solve it.

Lately though, we’ve been exploring problems and finding solutions that are straightforward but impossible to realize. The problems are two-fold: things that are too simple to make for meaningful research problems, and there is a barrier of platform lock.

For example, a team of GT students led by Hayley Evans found that people trapped in the economic crisis in Venezuela are increasingly using Facebook Groups for barter of basic necessities. It’s no longer possible to buy diapers at a fair price, but you can trade staples like flour for them. However, people are still price gouging and duping others with fake products. The solutions here are simple—a price comparison tool like the one Stubhub provides for ticket sales could give everyone a calibration on what exchanges are fair. A reputation system like the one on ebay could help stop scammers. If people have public reputations, then individuals can choose not to trade with someone who has a negative reputation and be extra careful with a new account with no history. These are established solutions, but it’s not clear who can build these tools for the Venezuelans, even though the need is desperate. It’s certainly not research—it’s too easy. To do something as a research project, we need something that we can raise grant funding for and publish about—we need to innovate. But solving this problem doesn’t necessarily need much innovation.

The second part of the problem is platform lock. Venezuelans are using the platform they are already on—Facebook. It would be hard to imagine bringing people to a custom platform, even if we had the time and resources to try to build one. And although you can make some small changes to platforms like Facebook with browser scripting, those solutions are limited and fragile.

Here’s another example. In 2012, Dimond did a study of the use of mobile and social computing by survivors of domestic violence. Her research concluded that there are some simple things that could really help people in this situation. For example, features developed as parental controls could be adapted to provide protection from harassment for adults. But we have the same problems again—the solutions are largely so simple as to not qualify as research, and they’d have to be implemented by mobile carriers. What is missing is societal—why can’t we find resources to do these simple things? A purely market-based model for software development falls short of meeting people’s needs.

These issues are going to multiply. As software reaches into more and more nooks and crannies of everyday life, we need an economic model that can deliver needed features that don’t make sense from a pure profit motivation. This will involve more activity by software nonprofits like Mozilla that design tools for the public good. It will further require better computer science education and extensible platforms so that people can develop solutions for themselves.

Advertisements
%d bloggers like this: