Local-first Software for Frontline Communities

Karissa McKelvey
No items found.
October 29, 2019

Cloud software simply doesn’t work where frontline communities live and where oil spills happen.So we built Mapeo differently, with a set of principles called “local-first software,” a term coined by Ink & Switch.These principles allow for both the collaboration and ownership of data entirely offline. This software works better, too -- no loading spinners, no broken links, and no data hostage.Local-first means that data is stored on the device (laptop or mobile) first. When people are ready to share the data to another device, Mapeo synchronizes the data with the device of choice: laptop, phone, raspberry pi, home or even cloud infrastructure. The point here is consent over where the data is synchronized.We couple this with free and open source principles that bake in freedom and autonomy. This improves the conditions for agency, long-term preservation, and user control, and has the potential to shift the dominant paradigm of how technology is built. We are pushing the boundaries of what’s possible in technology so that we can better stand in solidarity with marginalized communities.

How is this different than existing technologies that work offline?

Mapeo isn't the first mapping software that works offline. In fact, there's a long history of offline-first technologies designed specifically for mapping and monitoring. However, these differ because they are still clients in a hierarchical relationship with the cloud. The device is just holding data, waiting until it can connect to the cloud service.Local-first software, on the other hand, can synchronize with any desktop or mobile device without the cloud. These are peer to peer or distributed systems, where each device is running the same software as the next device. Each device is equal to another, with no hierarchical relationships between devices.

It’s faster software

Applications built this way are more responsive -- no waiting for cloud infrastructure in another city or country. This is even more crucial in remote at-risk areas, like the Amazon Rainforest or the Arctic. The Internet is either non-existent or too slow even for basic browsing, as connections fail before finishing the round trip between distant data centers across the world.

Server-to-server round-trip times between AWS data centers in various locations worldwide from Highly Available Transactions: Virtues and LimitationsBut won’t the Internet get faster? Is local-first software just a temporary measure until we build more satellites and lay more fiber cables? Unfortunately, as the Internet gets faster, software becomes more data intensive. We have seen this trend since the beginning of the Internet, as web-scale applications increase bandwidth requirements by orders of magnitude every few years. The local-first pioneers have a combined decades of experience working in remote areas, which caused them to double down on the bet that cloud software will always be too slow for these remote areas.And in many of the places we work, we want it to stay that way -- solidarity means protecting these pristine environments and supporting the people who defend them. In other words, it is more preferable and practical to utilize the collective data center that already exists inside of the devices around us -- instead of destroying more of the environment to lay cable and satellite infrastructure.

It’s more effective technology for good

Most software today collects as much information as possible about it’s users -- even if there isn’t a clear and immediate use. This sensitive user data is hoarded as a future asset by cloud software, even within the “tech for good” nonprofit technology movement. Teams utilize this data to improve the product over time, and many safeguard future profits by limiting data portability to another software product. To build and maintain this potentially massive data warehouse, organizations often get large rather quickly. Burning through millions of dollars without a revenue stream is justified by the future asset that is the data warehouse. This is the dominant business model established by Silicon Valley giants -- and "tech for good" organizations are largely following their lead.In local-first software, on the other hand, data is seen as a liability instead of an asset. In the case of Mapeo, frontline communities collect very sensitive data that needs to be held securely. Stories from indigenous elders, locations of sacred sites and herbal medicines, hunting paths, and photos of illegal mining, are sensitive pieces of knowledge that we don't want to get into the wrong hands. So for our technology team, an unencrypted data warehouse would be a huge undertaking with potential for negative consequences if done incorrectly. So we don't have one.By treating unencrypted partner data as a liability rather than an asset, we are able to keep the technology team small, lean, and agile. Rather than burning through cash to build a data warehouse, we can focus on what matters:reliable and efficient software, co-designed in person with users, that works anywhere. Instead of driving product decisions from metrics dashboards in a room thousands of miles away, we form direct partnerships and workside-by-side in collaborative design processes. The only data we collect in Mapeo are error messages, used for triaging fatal errors and application crashes.

It’s about solidarity

Reversing the climate catastrophe isn’t profitable, so companies have no incentive to create local-first software for frontline communities. This is why it’s ever more important for the public sector, nonprofits, and private foundations to invest in this approach.So far, this software has been funded by a pay-what-you-can model that isn’t based on holding data hostage but instead on providing software that works well. It is simply asking individuals, foundations, companies, and organizations to donate for continuing the improvement of Mapeo.Donations aren't always monetary. We are part of a vibrant open source ecosystem of volunteer developers who care about local-first and peer-to-peer software, which includes Peermaps, Dat Foundation, Hyperdivision, Ink & Switch, Development Seed, Mapbox, and many individual contributors. Together, we work in solidarity with marginalized communities and give them the technology they need without asking them for money.Join us by contributing to GitHub or donating online here. Contributions like these are what keeps us going.

Published by
Karissa McKelvey
Back to the Blog