Word counting with \w+ is inaccurate for non-English languages

[mdroidian]

Description
The current word-counting approach using a regex like \w+ is not sufficient for non-English languages. It fails on many scripts (e.g., CJK languages, languages without clear word boundaries, or those using combining characters).

This is not necessarily a request to change the current behavior, but rather a note for future reference in case users report inaccuracies.

Context / Rationale

• \w+ is heavily biased toward ASCII/Latin scripts.
• Even Unicode-aware regex approaches still break down across languages.
• The closest modern standard approach in JavaScript is Intl.Segmenter, which performs locale-aware text segmentation.

Reference:
https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Intl/Segmenter

Notes / Caveats

• Intl.Segmenter is better, but still not universally “correct” in all linguistic contexts.
• Any word-counting solution will involve tradeoffs depending on language, locale, and definition of “word.”

Suggested Action

• No immediate change required.
• Keep this issue as documentation / justification if word-count accuracy for non-English text becomes a concern in the future.