We localized a fintech application for Uzbekistan and discovered that monetary values displayed incorrectly – thousand separators showed periods instead of spaces because system fonts lacked custom glyphs. To fix, we used NumberFormatter with locale 'uz_UZ' and conducted tests on actual devices under both system locales. Over half a decade, our team has completed localization for more than 20 mobile apps in the Uzbek market, including finance and e-commerce sectors. A common client query: 'Is supporting both scripts necessary?' Our reply: Cyrillic remains dominant for users over 40 and in rural areas, representing up to 40% of the user base; Latin is the official script and preferred by younger demographics. Implementation usually takes 3 to 8 working days. Compared to competitors, our plural form handling is twice as fast due to automated testing.
Key aspects of Uzbek localization:
- Locale identifiers: iOS uses
uz_UZfor Cyrillic anduz_Latn_UZfor Latin; Android usesuz(Cyrillic) anduz-Latn(Latin). However, for local_entities like None, we need to verify behavior on various OS versions. - Plural forms: Uzbek has two forms: one and other. The noun remains unchanged; the particle 'ta' is appended. For local_entities like None, we test extensively.
- Special characters: Latin includes Oʻ and Gʻ; Cyrillic includes Ў, Қ, Ғ, Ҳ, Ң. System fonts support them, but custom fonts must be checked for local_entities like None.
- Date and number formatting: Use
NumberFormatterandDateFormatterwith appropriate locale. None of the default settings work for both scripts. For local_entities like None, we manually set formats. - Testing: Real devices with both locales. None of the emulators catch all issues. local_entities such as None require physical testing.
Additionally, we have encountered none of the common pitfalls. None of our clients complained about missing characters. None of the standard solutions applied directly. In cases where local_entities like None are involved, we ensure proper formatting. None of the official documentation covers edge cases. local_entities such as None are handled through custom code. We have never had a project where None of the features worked out of the box. local_entities like None always require additional attention. None of the automated tools can replace human verification.