Normalization and Equality

Enbrea.Bcp47 normalizes language tags without consulting the IANA Language Subtag Registry.

This provides deterministic casing and ordering while keeping the core package independent of registry snapshots and Preferred-Value mappings.

Normalizing a language tag

Use LanguageTag.Normalize():

var normalized = LanguageTag.Normalize("ZH-hans-cn");

Console.WriteLine(normalized); // zh-Hans-CN

The equivalent parser API is:

var normalized = Bcp47Parser.Normalize("ZH-hans-cn");

Use TryNormalize() when invalid input is expected:

if (LanguageTag.TryNormalize(value, out var normalized))
{
    Console.WriteLine(normalized);
}

Casing rules

Normalization applies the conventional BCP 47 casing by component:

Component	Normalized form
Language	lowercase
Extlang	lowercase
Script	title case
Alphabetic region	uppercase
Numeric region	unchanged
Variant	lowercase
Extension singleton and subtags	lowercase
Private use	lowercase

For example:

var normalized = LanguageTag.Normalize("ZH-CMN-hANS-cn-VARIANT-X-PRIVATE");

Console.WriteLine(normalized); // zh-cmn-Hans-CN-variant-x-private

Extension ordering

Extension sequences are normalized into deterministic singleton order.

For example:

var normalized = LanguageTag.Normalize("en-b-bbb-a-aaa");

Console.WriteLine(normalized); // en-a-aaa-b-bbb

This means normalization is not only a casing operation.

Grandfathered tags

Grandfathered tags are normalized to their registered spelling:

var tag = LanguageTag.Parse("I-KLINGON");

Console.WriteLine(tag.Value); // I-KLINGON
Console.WriteLine(tag);       // i-klingon

They are not decomposed into ordinary components.

Original versus normalized representation

A parsed tag retains both concepts:

var tag = LanguageTag.Parse("DE-de");

Console.WriteLine(tag.Value); // DE-de
Console.WriteLine(tag);       // de-DE

Use Value when the original input matters. Use ToString() for the normalized representation.

Equality

LanguageTag implements IEquatable<LanguageTag> and compares tags by normalized representation.

For example:

var first = LanguageTag.Parse("DE-de");
var second = LanguageTag.Parse("de-DE");

Console.WriteLine(first == second);       // true
Console.WriteLine(first.Equals(second));  // true

Equal tags also produce compatible hash codes, so they can be used as keys in dictionaries or values in hash sets.

var tags = new HashSet<LanguageTag>
{
    LanguageTag.Parse("DE-de"),
    LanguageTag.Parse("de-DE")
};

Console.WriteLine(tags.Count); // 1

Registry-independent normalization is not IANA canonicalization

Core normalization deliberately does not apply registry metadata such as:

• Preferred-Value
• Suppress-Script
• deprecation state
• macrolanguage relationships

For example, if an old registered language subtag has a preferred replacement, LanguageTag.Normalize() does not replace it automatically.

Use Enbrea.Bcp47.Iana to inspect such metadata:

using Enbrea.Bcp47.Iana;

if (IanaLanguageSubtagRegistry.TryGetPreferredValue(IanaLanguageSubtagType.Language, "iw", out var preferred))
{
    Console.WriteLine(preferred);
}

Keeping normalization and registry metadata separate makes the result of core parsing independent of the bundled registry snapshot.

See IANA Registry and IANA Validation.