Part of the challenge of designing this system seems to arise from the fact that it could actually serve a variety of purposes. As it turns out, these purposes can be arranged in a roughly hierarchical order with low-level purposes providing foundations of functionality for higher ones. This is an attempt to enumerate these and, for each one, clarify the unique engineering and design requirements and explore the advantages that Wicker might offer over competitor systems.
@todo: differentiate between “use cases” and “functionalities” (don’t assume that cool functionality equals value to users).
1. Knowledge management
One of the most foundational use cases that Wicker could have — and almost more of a system than a use case — this applies to fact-based knowledge and involves building a map of facts and sources. It has applications in fact checking, figuring out what sources are and are not saying, and problems related to the management of sources, references, and fact-based information in general, e.g., detecting circular references, ranking the strength of facts, and looking up information with highly traceable and reliable sources.
This could be interesting to people who:
- Are interested in contributing to a durable, public knowledge base
- Are interested in verifying the truth of claims
Other use cases that are supported by this include:
- Learning, especially if knowledge management tools are localized to specific regions or companies where competitor learning tools have less information (or less reliable information).
- Decision-making, since decision-making often requires reliable knowledge.
2. Learning
This is the umbrella case of exploring a topic. Compared to internet searches and chatbots, Wicker could (in principle) offer denser links to sources if it builds on a good knowledge management system. This is the most basic augmentation of a knowledge management system — it’s sort of just adding an query layer — though it may also take information from sources outside the KB.
3. Modelling reasoning
This is similar to knowledge management but it applies to how we think about things, including what we like and dislike, what we deem likely and unlikely, and how we draw conclusions about things. It includes specific processes such as examining tradeoffs. Here, in particular, we should be careful about confusing functionality with value to users. It may be possible to model reasoning, but unless it’s packaged with some kind of really satisfying user experience, there’s no point.
4. Communication
One of the key features missing from our current era of chatbots, the ability to contribute to a public conversation or knowledge base is, is the first layer in the hierarchy of use cases where Wicker starts to become worth building. It marks a minimum threshold for significant usefulness of Wicker — if it cannot be supported, then Wicker should not be built. Communication is where the modelling of debates can begin to take place. It also may foster the sense of being part of something. It builds on the foundations of knowledge management by adding a user input layer and solving all the attendant problems such as duplication of claims, the handling of meaningless claims, etc.
5. Cultural knowledge management
Survey-style stuff; pulse of public opinion. Builds on communication by storing and managing the opinion-based input in a structured way, including offering the ability to derive insights from it. This would also require augmentations to the underlying knowledge representation.
6. Decision making
Decision making can be further broken into personal for single users, collaborative which includes small groups and cases where collaborators are synchronously present, and mass-collaborative which operates with large, asynchronous user groups and overlaps significantly with “digital democracy” efforts. The collaborative cases build on the communication case since they require users to contribute information — probably both facts and opinions, and potentially even humour. Decision making includes some unique cognitive and social processes which would need to be modelled as distinct features in a system like Wicker:
- Voting
- Juries often form some narrative around their case
7. The unified system
The problem with breaking down the above use cases is that, in practice, you might actually want everything to happen in a single system, or at least for adjacent systems to be highly integrated. If you really care about doing something at scale, the system cannot be to purely academic — it has to be enjoyable to use. Our decision-making process is often inseparable from our communications, learning, and knowledge management processes.
The role of local
Systems that support any of the above use cases would be particularly useful in the context of local issues, whether they be local to a geographic region or municipality, or local to a company environment.