DemocracyOS team at work.The Politics of Code is an interview series curated by Marco Deseriis with software engineers and political activists about a new generation of decision-making software that allow movements, organizations, and parties to make collective decisions online. Rather than reducing technology to the status of a tool, the interviews explore the different conceptions of political participation and democracy embedded in each software, the relationship between online and offline deliberation, as well as questions of authentication, verifiability, ownership, trust, and leadership.

Launched in 2013, DemocracyOS is a decision-making software for political consultation, deliberation, and decision-making developed by a team of Argentinean programmers and activists. The software was originally designed to support the organization and decision-making of the Partido de la Red, a small “Internet party” founded in Buenos Aires in 2012 and inspired by the model of the Swedish and German Pirate Parties. Over the past two years, the software has spread to several Latin American countries, Europe, Africa, and North America. It has been used by the Revolución Democrática party in Chile, Podemos in Spain, the national governments of Mexico and Argentina, the City of Nanterre, the D-Cent program funded by the European Commission, as well as in dozens of other experiments.

At first sight, DemocracyOS is less sophisticated than other decision-making software such as LiquidFeedback and AdHocracy, which allow participants to delegate their vote to other users and rank proposals in order of preference. However, the apparent simplicity of DemocracyOS is deceptive, as the software allows for a high level of customization and adaptability.  In this interview, Guido Vilariño, chief technology officer of DemocracyOS and founding member of the Partido de La Red, explains why a usable software interface is inherently democratic, how the community of DemocracyOS users drives the development process, and how online participatory democracy is a learning process for institutions and citizens alike.

Marco Deseriis (MC): Can you give us a short history of DemocracyOS? Why did you decide to develop this software?

Guido Vilarino (GV): We initially developed the software to support participation in the Partido de la Red. Our political proposal was simple: our representatives would always vote according to what citizens decided on the online platform.

After studying other decision-making software, we realized that they did not fit our needs so we thought it would be best to develop our own and share it with the rest of the world. Eventually, what started as a software tool to support participation in the Partido de la Red took on a life of its own. We renamed the software DemocracyOS, and made it more generic and easy to implement by others. Thanks to the support of various grants and philanthropic organisations we were able to build a development team that now employs two full-time software engineers, myself as a software director and engineer, and hired contractors. That’s just the technical team. The broader DemocracyOS project includes six more people, including anthropologists, educators, sociologists, and other professionals.

MD: What are the principles that undergird the conception and design of DemocracyOS?

GV: I would say that the development of DemocracyOS has been inspired by three fundamental principles: simplicity, openness, and modifiability. The implementation of these principles in the user interface (UI) and user experience (UX) determines the way users access information, the way they interact with one another and their elected representatives, as well as the adaptability of the platform to different sociopolitical contexts.

Screenshot with a pie chart of a voting result.

MD: Let us begin with simplicity. In what way is DemocracyOS “simple”?

GV: DemocracyOS is simple because it offers just a few features. Since democracy is not the privileged domain of the technologically minded, we have designed the platform to make it accessible to as many people as possible. Given the wide diffusion of smartphones, we decided almost immediately that DemocracyOS was going to be a mobile-first application. Thus we had to work within certain boundaries as there is only so much you can do on a cellphone screen.

There is only so much you can do on a cellphone screen.

Following the principles of lean design (a design method that focuses on eliminating unnecessary tasks and redundant information from the user experience, Ed.) we decided that proposals could be summarized in a few sentences and that users could express arguments in favor of or against them immediately below this initial summary. However, each summary can be deepened through links to the extended proposal, possibly with side notes, which become visible only if the user is interested in reading them.  

To introduce these features we did not have to reinvent the wheel. We took inspiration from a number of existing applications such as Facebook, Quora, and StackOverflow, and we were able to implement the comment feature in two weeks.

MD: Can users reply to each other? Do discussions branch out in subthreads?

GV: They can, but in a simple way. A user may respond to another user, and a subthread emerges from this interaction, but that is the only conversation depth there is. You can find this kind of feature on Facebook for example.

Conversely, on platforms such as Reddit (which we also find quite interesting), each response can spin out a new thread open to everyone. Yet multithreaded conversations can become very difficult to follow, especially when complex ideas are debated. And of course they are not very cellphone-friendly.

So, from a user perspective, a multi-threaded approach is really demanding in terms of time and effort you must put into the app as opposed to just focusing on the issue at hand. It is also an activity that ends up privileging those who have more time on their hands, and that isn’t very democratic. Ultimately, technology does its job well when you do not realise it is there and you can just do what you want to do, without bothering about buttons or scrolling.

MD: So you are saying that software affordances that are simple and easy to use allow for a leveling of the playing field in terms of democratic participation. I get that, but democracy in a modern society is also a complex affair, which involves multiple actors that interact with one another on so many levels.

Don’t you think that compressing all these layers in an app is at risk of reducing democracy to a prepackaged template that does not account for the manifold dimensions of democratic participation, deliberation, and decision-making?

GV: No I don’t. To begin with, DemocracyOS does not pretend to be the only and definitive platform for every aspect of democratic life. Second, DemocracyOS is not “prepackaged” in that it is a highly flexible tool that can be adapted to many different uses and circumstances. DemocracyOS is Free Software (Free as in “freedom,” not “gratis” – also known as Libre Software), and is released under the third version of the GNU General Public License (GPLv3). The fact that our software is open source ensures that everyone can look into the code, which is of course a question of democratic control. Further, because the software is freely redistributable the community of users benefits as a whole from each software modification.

Another variation of the DemocracyOS tool.

MD: Why did you opt for the GNU license as opposed to an open source license that is not free?

GV: In hindsight, we could have used a more programmer-friendly license such as the MIT license. But under the MIT license, which is also open source, developers are free to take the software and do what they please with it, including releasing modified versions as proprietary software, without having to disclose how they made certain changes and why. A GPL license instead puts the entire community of users at the center of the development process in that any derivative version of the software must be released under the same license. While this makes it a little more bureaucratic for developers, it protects users’ rights as it guarantees they will never be using a version of the software that cannot be audited or redistributed for free. Javascript allows us to reach a community of potential co-developers that goes well beyond the professional programmers.

MD: So it is as if you have two concentric circles, the circle of developers and the circle of users, with the former working in the service of the latter. It all sounds very democratic, but developers have programming skills and Internet users for the most part do not...

GV: Yes, of course, and this is the reason why we decided to develop the platform in Javascript. If you can write code and have worked on anything that has to do even faintly with the internet, it is quite likely that you will have at least basic knowledge of Javascript. This is not something we just believe in, there is statistical evidence to back this claim: Javascript is by far the most popular language on Github (a large repository of open source software, Ed.) and the trend does not seem to be receding. Thus Javascript allows us to reach a community of potential co-developers that goes well beyond the professional programmers. In this sense, the simplicity and usability of the user interface has a counterpart in the accessibility of the source code, which is really quite easy to read and modify. I also want to stress that the principle of modifiability has immediate political consequences in that it allows for the adaptation of the platform to different democratic practices rather than the other way around.

MD: Still, you need someone in your community with basic programming skills to make these modifications.

GV: Yes and no. DemocracyOS is already designed to allow for a certain level of customization. For example, if you start a forum you can set it to be open to all registered DemocracyOS users, where everyone can comment and vote. Alternatively the forum can be restricted, allowing everyone to see the topics but only some people to comment and vote.

A third option is that the forum is set to be secret, where nobody can see the topics or vote unless they are authorized by the owner of the forum. For example, most discussions within the Partido de la Red are restricted. Everyone can see what we discuss but only party members can vote. Additionally, platform administrators can tailor how voting works, how many options are available. Or they can disable voting altogether when all that is needed is a fruitful discussion. Everyone can see what we discuss but only party members can vote.

MD: If participants feel they are not enough competent on a given issue to express a vote, can they delegate their vote to a person they trust?

GV: Not yet, but we are working on implementing this feature.

MD: Aren’t you concerned that in countries like Argentina where the level of perceived corruption is quite elevated, proxy votes could be used for vote-buying and corrupt the political process?

GV: Let me start by saying that corruption and vote buying are sociopolitical issues that are beyond solving by any software tool. These kind of practices flourish especially when they go unaccounted for as corruption works best when few people engage in the discussion. This is why an effective way to fight against these practices is to demand and implement more transparency.

The fact that DemocracyOS makes all debates and votes visible to everyone helps citizens to be more responsible and accountable for how and why they participate, as is the case with existing offline public hearings. The possibility of delegating a trusted person on specific issues can only make this assumption of responsibility more explicit, which is a good thing in my opinion.

MD: Is it mandatory for DemocracyOS users to use their legal names or can they use nicknames?

GV: We encourage people to participate with their real names. However, we understand that in certain circumstances anonymity is necessary. For example, we were contacted by a person in Kenya who wanted to use DemocracyOS but feared for their safety. So we helped them set up an anonymous account.

In an almost opposite scenario, a community wanted their users to be able to register with their Facebook accounts so they could easily learn who was who on the platform. In this case, we implemented a Facebook login and a Twitter login. We try to meet the needs of different communities to make it as easy as possible for them to use the app.

MD: Yet is seems to me that the overall logic of DemocracyOS is to make decisions transparent and visible to everyone. What is the rationale behind this? The secret ballot is still an essential property of the democratic process to protect voters from intimidation and other forms of abuse...

GV: I would say that the primary reason for making individual decisions transparent is technical. Because a secret electronic vote could be easily manipulated, we ensure verifiability by making every vote is visible to all participants. Yet we also try to limit undue influence on the final outcome of a ballot initiative by showing the results only after the discussion has ended and all the votes have been cast. In this way, even though all users are free to express their opinion in favor of or against a proposal, they can change their mind until the very end.

MD: Still, a technical limitation has a significant political effect in that it forces users to be always public about their decisions. Whereas a local debate about how to allocate city funds should be relatively free of potential intimidations (although this may not be the case in areas where organized crime and corruption are high) there are many issues on which citizens should be able to cast a vote without fear of reprisal.

GV: There may be issues for which voter privacy may be needed. That being said, I want to point out the enormous difference between discussing and voting on a piece of legislation or an issue of public interest. Whereas voting often requires privacy, discussing legislation is commonly a public activity in modern democracies.

At this stage, we see DemocracyOS mainly as a platform that helps build a more informed citizenship and can make institutions and elected representatives more accountable. But of course we believe that it will work best when citizens’ choices will become binding. Still, civic participation is very much like any other human activity, you learn by doing it. Citizen participation was higher than expected, and the government closed the consultation.

MD: Give us an example of how this process can be learned?

GV: Sure, DemocracyOS was recently used for a government consultative initiative on the future of public education in Argentina. The Presidency of the Republic invited all Argentines to express their opinion on the Declaration of Purmamarca, a programmatic document drafted in February 2016 that aims at redesigning the primary and secondary systems of education.

Citizen participation was higher than expected, and the government closed the consultation without taking any concrete action or communicating their next steps. This implies that even if governments are willing to open up to citizen input, that openness must include a clear, concrete follow-on process, a way to apply that input to real policy and actions.

If not, citizens lack the feedback loop that encourages sustained participation and the well-intended effort can lead to even more apathy. Perhaps the government wasn’t ready for that much input and so didn’t design a follow-on. But this wouldn’t have been clear without the consultation.

This is what I mean by a learning process: governments have been endogamic institutions since forever and they aren't used to opening up, but there is some political will for it to happen, and we'll see a lot of trial and error in the process. Still, any move forward is progress, and we need to celebrate it.

MD: I see. What would you do if the government should fall short of setting up such a process? Are you offering any advice to them?

GV: Yes, we provide a lot of assistance on how to set up follow-on strategies, processes, which roles should be there to handle citizen input, and so forth. We're happy to help, but in the end it is up to the institution to choose how to handle it.

MD: Finally, I would like to close this interview with a political question. Critics of open data and open source “ideology” argue that informational capitalism thrives on participation and transparency.

Platforms such as DemocracyOS, Loomio, LiquidFeedback and the whole #civictech “movement” encourage citizens to participate in civic life and express their political opinion in the open. On the one hand, civic participation can reduce administrative costs as local governments can outsource some services to the citizens themselves. On the other hand, these platforms could provide the data mining industry with a wealth of data on their political preferences that can be easily correlated with consumer preferences.

In what way does DemocracyOS guarantee that the data citizens generate through political participation are not taken out of context and used for business purposes? In other words, is there a way of decoupling open data from data mining?

Is there a way of decoupling open data from data mining? GV: Great question.

GV: Great question. The way I see it, capitalist actors will always attempt to make a profit out of every opportunity, which isn’t necessarily good or bad, it is just the way it is. As with the issue of corruption we discussed before, this is beyond DemocracyOS or any other software tool.

The key for us is to provide the same data to everyone so that independent activist organisations and NGOs can have access to as much data as big corporations and other powerful institutions. In the past, only the owners of large capital investments could muster the effort of performing large scale consultations or polls, which left smaller actors in a position of disadvantage.

By providing the same level of access to anyone, we are trying to level the playing field. This allows smaller actors to step in and, as the history of the Internet has taught us, it ends up democratising information, especially since smaller organisations can move faster than big ones. Smaller organisations can move faster than big ones.

Big industry and state actors are already collecting and browsing all kinds of information about citizens. The way to fight back is to provide citizens with tools that have the same capabilities but in an easy-to-use fashion. That said, the whole state surveillance apparatus that we have been learning more and more about lately is of great concern to us, which is why we provide free and open tools and guides for easily encrypting instances of DemocracyOS and similar software platforms.