Help us complete the word counting algorithm!

[Josh-Cena]

Word counting is not as trivial as it seems. Most languages use spaces to separate words, but languages like Chinese and Japanese treat each character as a word, and there are no spaces. Therefore, we treat them slightly differently from Latin languages.

If you:

1. Know a non-CJK language that doesn't use spaces to separate words
2. Know the exact unicode range of characters that can be "standalone words"
3. Have backed up arguments about whether Katakana characters should be counted as words (see feat: add CJK support #35 (comment))
4. Other valuable comments about our implementation

Please don't hesitate to open an issue, send a pull request, or reach out to me or @ngryman in any possible way!

[EatChangmyeong]

reading-time/src/reading-time.ts

Lines 74 to 83 in 1d07a5c

	// A CJK character is a always word;
	// A non-word bound followed by a word bound / CJK is the end of a word.
	if (
	isCJK(normalizedText[i]) ||
	(!isWordBound(normalizedText[i]) &&
	(isWordBound(normalizedText[i + 1]) || isCJK(normalizedText[i + 1]))
	)
	) {
	words++
	}

Hangul characters are not words on their own, rather it works like English and separate words by space (e.g. "다람쥐 헌 쳇바퀴에 타고파" is only 4 words, not 11; actually postpositions complicate things but counting spaces is good enough for reading time estimation). I don't know about Chinese very much, but AFAIK Japanese will require a sophisticated algorithm because it doesn't use spaces but still has multicharacter words.

I found this package through Astro's docs and find it useful, but this makes me use an absurd reading speed of 500 WPM as a workaround.