Walter is a prototype for a global, decentralized, question and answer framework with a built in compensation mechanism so researchers can be compensated by contributing work.
Platforms like Wikipedia have shown us that people are willing to contribute significant time to expanding and recording shared knowledge. The advent of decentralized applications (blockchain technology or Web 3.0) enable us to take the concepts of building shared knowledge and doing collaborative research significantly further.
Codename, as a founding member of Orbit Technologies, developed the underlying theoretical premise of the project, co-authored the white paper, designed and lead product development through to a working prototype of an ERC-20 compliant token on the Rinkeby testnet.
The Walter prototype was a first look by our team into the Ethereum blockchain to see if it was a suitable platform for a trustless, fully decentralized and open source global research protocol. Blockchain offers the promise of a distributed community everyone would know the rules, incentives would be aligned and rewards would be provably fair and enforced via smart contracts.
Our goal was to create an ecosystem where data and knowledge could be exchanged but also collaboratively generated. All data submitted would be made permanently available to anyone in the community and the results would have to be immutably recorded, open to all and the research platform could not be manipulated by a third party.
What emerged was a token-backed two-sided marketplace platform consisting of actors who request data (Consumers) and others who supply responses to the requests (Producers).
At the core of the platform is the ERC-20 compliant “Walter Token”. It is the internal currency with which Consumers can request data and subscribe to ongoing feeds and the mechanism for compensating Providers for supplying responses. Any member of the community can be both a consumer or producer (though not on the same question). In addition a separate reputation score is also written to the blockchain for all actors. Reputation is used in proprietary ways to manage what are known as Sybil attacks.
The greater goal was to move to a decentralized knowledge creation and data exchange where data is created open and value accrues to the network as a whole – that is as the data created within the community gains value so too does the underlying token.
We’re at the very early stages of blockchain technology and it is evolving rapidly. Those two factors combined (early stage, changing quickly) make for interesting design and user experience challenges. We tried to mitigate this instability with two perspectives on working with exponential technologies.
First, find what has been built, cobble it together and work ugly. Much of the tooling that’s required is very early in this development, or more often has not yet been built. There are some fundamental tools being developed but providing a coherent experience in a new ecosystem can be challenging, we tried to get the ugly stuff right before putting a face to it. In other words use what’s there without getting tied to it.
Second, if the technology is in flux the underlying experience of the application needs to be sufficiently familiar to users up until they hit the complexity that the blockchain in its raw state introduces to a project. To address this we based the question and answer experience on many popular apps that push a set of questions to the user on set intervals and prompt users with notifications that a new round of questions is starting.
Beyond the core experience of having questions sent to Producers we also had to address two other key experience points. First, we worked to clearly guiding people through their set up, making sure to clearly explain the concepts of staking answers, and reputation.
Second we focussed on the answer submission process. In a web 2.0 environment this would have been a fairly simple set of patterns. In the case of the Ethereum blockchain even after looking at clever technical ways to remove lag on the front end, we hit our greatest challenge in managing delays in response time, which range from seconds to minutes and the resulting error handling. Even for a prototype / MVP this required special attention to confirmation messages and micro copy to manage a range of possible response behaviours from the systems.
The MVP was a success and a potential larger project for the next crypto wave.