Skip to content

Design Revisions

  1. Mindmap Storage and Edit: in the beta release, we tried to save the whole mindmap, including all the nodes and edges, as a stringified JSON in the backend. In the convergent design phase, the backend design of the Mindmap concept is more complex. We chose this approach for the following reasons:

    • follow the original data structure of the vue-flow library
    • fast implementation under a tight time budget
    • clear data storage and load logic, easy to manipulate through frontend operations

    There are also drawbacks:

    • Potentially lower performance for complex mindmaps
    • The nodes and the edges are at the same level, making it difficult and slow to filter by iterating through objects
  2. Functionality and Usability

    • Given the time budget, we prioritized the implementation of our core functionalities: mindmap, idea cards, and autosuggestions. The discussion concept we may implement in the next iteration.
    • Between generating nodes and generating connections, we prioritized the connections. We believe the finding connections between ideas can be the most time-consuming part of the user journey, and meaningful connections will generate more inspiration for the users.
    • As the wireframe depicts, we are trying to make the app as clean and functional as possible. In the actual implementation, we removed the separated note creation page and merged it into the note list. The user won't have to jump to another page to create new nodes, which streamlines the user journey.
  3. The usage of the Large Language Model as a suggestion engine

    • In the actual implementation, the usage of OpenAI API sometimes provides unstable results. It may be solved with more precise prompt engineering.
    • In the future, we may also try out more advanced NLP libraries that work in the backend to provide better suggestion methods.
  4. Design changes' influences on impact cases

    • Current file-based Mindmap data structure may become an obstacle when the growing user base generates complex mindmaps. The low performance will reduce user experience and reduce the impact on social influences.
    • Simplified collaboration function may reduce the impact case of collaborative brainstorming. It would be better if real-time collaboration, like miro board, could be implemented in the future
    • The creation assistant functions can be more powerful in multiple ways to generate more social impacts, including providing graphic and sound feedback, generative nodes, and grouping ideas.

Plan for User Tests

Task List

Feedback and Discussions Questions

  1. Tell me your worst experiences in this test.
  2. What do you think about the color scheme?
  3. What do you think about the font size?
  4. Tell me what left the most impression on you.
  5. Any suggestions on refining the app?