18-03-2026, 03:23 PM
Thanks — that’s a really helpful clarification.
If I understand correctly, you're pointing out that any apparent “roles” might depend heavily on how tokens are segmented in EVA, and that something like “daiin” may not be a stable unit.
That makes sense, and it probably means that before looking for functional patterns, I should first consider whether the units themselves are well-defined.
Would you say that testing patterns at a more granular level (e.g. subcomponents like “iin”) is more appropriate, or is the ambiguity too high even there?
I’m trying to understand where a structural approach would even begin to be meaningful.
If I understand correctly, you're pointing out that any apparent “roles” might depend heavily on how tokens are segmented in EVA, and that something like “daiin” may not be a stable unit.
That makes sense, and it probably means that before looking for functional patterns, I should first consider whether the units themselves are well-defined.
Would you say that testing patterns at a more granular level (e.g. subcomponents like “iin”) is more appropriate, or is the ambiguity too high even there?
I’m trying to understand where a structural approach would even begin to be meaningful.