Suggest modification:
If title has >50% letters which are non-ASCII (and non-ASCII Latin Extended, which basically means code point > 255)…
Turn off rules related to English (e.g. complete sentence, all-caps etc.)
Because under such circumstances, it is highly likely that:
- The title is not in English
- The English letters used in the title are probably ID’s, model numbers, codes (e.g. stock code), abbrev, prefix, etc. that should be left as-is
- Applying English styling rules is a mistake
This can be simply done by:
- Define
non-englishif (ch > 255) / (Latin letters total count) > 1 - Do not apply
prettifyonnon-english - Do not apply
seems_quietonnon-english - Do not apply
seems_pronounceableandseems_unpretentiousonnon-englishbecause their punctuation may not be in the ASCII set (e.g. CJK punctuation are all > 255)