Locale behavior, file quality, layout fit, and what a round costs
Four utilities that compute from live Intl data and the browser's own text metrics. The explorer and the estimator keep their state in the URL, so a configured view is a link.
CLDR explorer
Find which shipped locale breaks a date, number, or price.
One value formatted across every locale in a set: field order, separators, digits, sign position, and the invisible characters that break parsing. Exports as Markdown or CSV.
Writing specs, filing formatting bugs, checking whether it is a bug or a locale
String audit
Check a file before it goes out, and a delivery before it comes in.
A structural lint for source catalogs, plus a bilingual gate for returned deliveries: missing keys, dropped placeholders, and the plural arms the target locale requires.
Vendor handoffs and pre-import checks
Text expansion
See how much wider a string gets before it reaches a translator.
One UI string machine-translated into 55 languages, then measured: characters, pixel width in a font the reader picks, and the ink height that clips Thai and Burmese in a control sized to Latin.
Sizing controls, writing a layout spec, reviewing a design
Budget estimator
Price a round per locale, and see how long it runs.
Word count, tier, and locale set in; per-locale cost and timeline out. Applies per-pair rate indexes, translation-memory leverage, and one-time engineering setup.
Scoping a launch, or a first pass with finance
The background, when a result surprises
The chapters behind these tools: formatters on why the explorer never hardcodes a pattern, pseudolocalization on the checks the audit runs before a file goes out, and the translation pipeline on the estimator's word counts and tiers.