vTaiwan was Audrey Tang’s first stab at designing a participation space that would connect Taiwan’s online generation with the nuts-and-bolts of government policy making. vTaiwan, to date, has no constitutionally binding authority over government legislation, but since its creation it has been deployed dozens of times to understand and elicit public opinion on issues including the regulation of Uber, online alcohol sales, and the creation of what Tang calls a “FinTech sandbox”—a scheme that allows companies to experiment with financial products that are not technically legally under current regulations for a limited period of time.

A similar initiative, called Join, which is fully government-run and also overseen by Tang, includes in its ambit every aspect of government action and has registered 10.5 million unique visitors. In a nation of 23 million, that’s pretty decent click-through.

Both Join and vTaiwan are built on top of Pol.is, an open source software program best described as a mechanism for developing consensus on disputed issues. “Pol.is,” says cofounder Colin Megill, “is a tool for turning crowds into coherence.”

Megill’s operating theory is that party politics in Western democracies is predicated on the exploitation of “wedge issues” to divide the electorate. Megill believes that “new computational methods” can be deployed to find areas of consensus, rather than division. Po.lis, he says, “gives agenda-setting power back to the public itself.”

Pol.is is intended to be an antidote to the polarization nurtured by traditional internet discourse. If Tang is a person one can’t imagine being in a flame war, then Pol.is is a program purposely built to prevent flame wars. “There’s a lot of very intentional design that makes sure that people can only add to, but not subtract or detract from the conversation,” Tang says.

Her favorite example: There are no “reply buttons” in Pol.is. All you can do is agree or disagree with a statement about a given topic (say, should Uber be allowed to undercut established taxi companies on price?).

Reply buttons, Tang says, are an invitation to trolls to wreak havoc by spreading disinformation, engaging in invective, or creating distraction. If the interface restricts engagement to merely expressing approval or disapproval, the trolls lose interest, Tang says.

In Pol.is, success is defined by the achievement of clusters of agreement. The goal, Tang says, is not unanimity, but rather a concept borrowed from the open source software developer community: “rough consensus.”

“[Rough consensus]” is not that strong,” Tang says. “It’s just something [programmers] can live with, then go back and write some running code, and stop debating. That kind of rough consensus is the key in Taiwanese norm shaping, because it enables people to not squander their time on getting the fine consensus out but rather to agree on something that we can all live with. That is something that politics can learn from internet governance: If we can all live with it maybe that’s good enough. Maybe we don’t need everybody to be literally on the same side.”

Megill says Tang and CL Kao, a cofounder of g0v and former business collaborator with Tang, convinced him to open source Pol.is. Taiwan, he says, has polished the software to its “most complete example.”

“Without someone who wants to bring deliberative practices into government” Pol.is is just a hammer, Megill says. “Audrey is the carpenter.”

But she’s far from the only assiduous tool-user in Taiwan. “In terms of citizen-led, civil society engagement with technology for enhancing the democratic good,” says ITFT’s Monaco, “Taiwan is the most lively civic tech sector on earth.”

But how exactly did that happen?