Understanding the ISO 639 parts

ISO 639 is a multi-part standard for representing the names of languages. Each part covers a different scope and code form, and the Language package exposes a distinct type for each. Knowing which part you need is the key to choosing the right type.

The four parts

Part	Code form	Scope	Type
Part 1	Alpha-2	The most widely used languages	`Iso639Part1Language`
Part 2	Alpha-3	Languages and a number of collections	`Iso639Part2TLanguage`, `Iso639Part2BLanguage`
Part 3	Alpha-3	Comprehensive coverage of individual languages	`Iso639Part3Language`
Part 5	Alpha-3	Language families and groups	`Iso639Part5LanguageFamily`

Part 1 — Alpha-2 codes

Part 1 assigns a two-letter code (for example en for English) to the most commonly encountered languages. It is the smallest of the code sets and is what most people picture when they think of a “language code”.

Part 2 — Alpha-3 codes, terminological and bibliographic

Part 2 uses three-letter codes. For historical reasons it was published in two forms, and the package provides a separate type for each:

Terminological (T) — Iso639Part2TLanguage. The code derived from the language’s own name (for example deu for German).
Bibliographic (B) — Iso639Part2BLanguage. The code historically used in bibliographic contexts, often derived from an English or French name (for example ger for German).

For most languages the T and B codes are identical; they differ only for the relatively small set of languages where the two naming traditions diverged.

Part 3 — comprehensive Alpha-3 codes

Part 3 extends the Alpha-3 approach of Part 2 to cover individual languages comprehensively, including many that have no Part 1 or Part 2 code at all (for example pht for Phu Thai). It carries the richest metadata of any part — scope, language type, print and inverted names, and macrolanguage relationships.

Part 5 — language families and groups

Part 5 provides Alpha-3 codes for language families and groups rather than individual languages (for example cau for the Caucasian languages). Its type is therefore named Iso639Part5LanguageFamily.

Why these are `IStringEnum` structs, not C# enums

A C# enum is backed by an integer. ISO 639 does not define a numeric value for its language codes — the code is the string. To model this faithfully, the Part 1, 2 (T and B), 3 and 5 types are not enums but readonly struct types implementing the package’s IStringEnum interface. Each language code is a static member whose underlying value is the string code from the standard, and the struct defines an implicit conversion to string so a code can be used directly where a string is expected.

This is why, in the how-to guides, codes are accessed as static members (Iso639Part1Language.en) just like enum members, even though the types are not enums. By contrast, ISO 15924 does define numeric codes for scripts, so Iso15924Script is a conventional C# enum.