Uzbek in two scripts
The official script is Latin, yet Cyrillic is widely used in documents and databases. The same person may be recorded in either alphabet, so two-way transliteration and storing both spellings as search indexes are required.
Operating on five markets is not a matter of translating the interface. Data residency, the identifier model, deadline calculation, the set of sources and document templates all change.
Some jurisdictions require personal data of their citizens to be processed in databases located on their territory. Retrofitting a system with the notion of a region after it was built without one means rewriting the whole data access layer.
It is set when the organization is created and does not change without a dedicated migration procedure.
Exchange between products of the line happens within the region of the organization.
A request reaches the correct environment before any business logic runs.
Backups are stored where the primary database is and are not moved between environments.
Information about legal requirements is provided as a design reference, not as a legal opinion. Before launching in a specific country, every point is confirmed by lawyers of that jurisdiction: requirements in this area change often.
An identifier is not a fixed-format field but a typed record. A subject may hold several, including from different countries: a citizen of Kyrgyzstan holding a Kazakh identifier is common in labour migration.
{ country, type, value }
Three regional traits affect not translation but search, string comparison and deduplication.
The official script is Latin, yet Cyrillic is widely used in documents and databases. The same person may be recorded in either alphabet, so two-way transliteration and storing both spellings as search indexes are required.
The transition is gradual and stretched over time. Names will appear in both scripts throughout the life of the product, so the transliteration mechanism is built in from the start as a shared component.
The alphabet has additional letters and its own collation rules. Comparing Mongolian strings under a Russian locale yields the wrong order, so sorting runs by locale rather than byte by byte.
Interface language is a user setting. Country, time zone, currency and calendar belong to the organization and are not derived from language: a Russian-speaking user in Mongolia is an ordinary case.
Source coverage differs between countries. Conflating these states creates a false sense of verification and is the most dangerous defect in systems of this class, so the platform separates them explicitly.
The source responded and the trait is absent. The result feeds scoring and is stored with a validity date.
The check could not be performed. Scoring does not treat this as a clean result, and the record shows the reason and the time of the last attempt.
We will tell you which sources are already connected in your country, what runs as stubs, and which points need legal confirmation before launch.