In considering the next version of Regroup, I tried to use Keicho and it didnā€™t work so well.

Why is that?

  • Keichoā€™s current priority is to delve into the keywords contained in the previous statement because ā€œI donā€™t feel like Iā€™m connecting the story if I keep coming up with quoted questions.
  • The result is inevitably a ā€œdig deepā€ style conversation on the topic

Width Preferred Keicho

  • Let the first question be level 0.
  • The level of the keyword that first appears in the answer to a question at level N is N+1
  • Make the system so that the score gets smaller as the level increases.

digression

  • If you record not only the level but also ā€œwhich children for which keywordsā€, you can create a tree of keyword development.
    • In other words, you can mind-map when you talk in a chat room.
  • Even with the current system, which isnā€™t width-first, you can import it into Regroup in a tree-like fashion if you keep track of the parenthetical statements.
    • That seems more beneficial.

This page is auto-translated from /nishio/幅å„Ŗ先Keicho using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. Iā€™m very happy to spread my thought to non-Japanese readers.