Thesis: Nextvi is actually more customizable than Vim or Neovim—not in toggle-on-a-feature time, but in creating something no plugin ever could, because the plugin boundary is itself a ceiling. Writing in C has no ceiling. And 7.5K LOC is easier to own than 300K.
Antithesis: But trade it against: maximum expressive power with maximum maintenance burden, vs. bounded expressiveness with stable, fast iteration. Raw capability wins on first principles. Engineering economics wins on sustainability. Neither is wrong; they optimize for different axes.
Refinement: So Nextvi's experiment is something not commonly done because it's not the path of least resistance. Maybe that constraint is the feature - it forces every addition to earn its place in C. No 50-line Lua script for a mediocre feature; it has to be 50 lines of C that don't break anything. That pressure keeps it at 7.5K instead of 300K.
Doubt: But doesn't that mean Nextvi is less extensible in the short term while more capable of radical evolution in the long term? A Lua plugin can't rewrite how the editor fundamentally works. A C patch can—it just hurts more.
Conclusion-as-open-question: So the real experiment is: what happens when you bet that the entire editor can stay small enough for one person to modify the core? The answer is either "beautiful coherence" or "ossification because nobody forks C for a minor feature." Most projects choose the plugin path not because it's better, but because it's survivable. Nextvi survives precisely because it doesn't. Whether that scales past one maintainer's lifetime is the open question.
Thesis: Nextvi is actually more customizable than Vim or Neovim—not in toggle-on-a-feature time, but in creating something no plugin ever could, because the plugin boundary is itself a ceiling. Writing in C has no ceiling. And 7.5K LOC is easier to own than 300K.
Antithesis: But trade it against: maximum expressive power with maximum maintenance burden, vs. bounded expressiveness with stable, fast iteration. Raw capability wins on first principles. Engineering economics wins on sustainability. Neither is wrong; they optimize for different axes.
Refinement: So Nextvi's experiment is something not commonly done because it's not the path of least resistance. Maybe that constraint is the feature - it forces every addition to earn its place in C. No 50-line Lua script for a mediocre feature; it has to be 50 lines of C that don't break anything. That pressure keeps it at 7.5K instead of 300K.
Doubt: But doesn't that mean Nextvi is less extensible in the short term while more capable of radical evolution in the long term? A Lua plugin can't rewrite how the editor fundamentally works. A C patch can—it just hurts more.
Conclusion-as-open-question: So the real experiment is: what happens when you bet that the entire editor can stay small enough for one person to modify the core? The answer is either "beautiful coherence" or "ossification because nobody forks C for a minor feature." Most projects choose the plugin path not because it's better, but because it's survivable. Nextvi survives precisely because it doesn't. Whether that scales past one maintainer's lifetime is the open question.