Home - Ziggy Data Language

Notation designed to help users grok data layouts

Structs vs Dictionaries

Ziggy has two notations for key-value pairs. Structs are for when valid keys are specified in a Ziggy Schema, and dictionaries are for when key names are up to the user.

This helps users get a better intuition for the schema before they even see it, and it also offers affodances like knowing when it makes sense to “goto definition” on a field.

Tagged Unions

Ever had to fight JSON to express a tagged union? Ziggy structs have names that can be used by both humans and tooling to discriminate between union cases.

…and more quality of life features!

Optional top-level curlies, multi-line string literals, trailing commas, comments, underscores in numeric literals, etc.

See an example package.json vs a (hypothetical) Ziggy version.

package.json

{
  "name": "ziggy",
  "version": "1.0.0",
  "license": "MIT",
  "dependencies": {
    "react": "next",
    "leftpad": "^1.0.0"
  },
  "repository": {
    "type": "git",
    "url" : "https://github.com"
 },
  "description": 
    "# Ziggy\n\nA Data Serializa..."



}

package.ziggy

{
  .name = "ziggy",
  .version = "1.0.0",
  .license = .spdx("MIT"),
  .dependencies = {
    "react": "next",
    "leftpad": "^1.0.0",
  },
  .repository = .git(.{
    // "type" is now the union case
    .url = "https://github.com",
  }),
  .description = 
    \\# Ziggy
    \\
    \\A Data Serialization Language.
  ,
}

A schema language for creating and documenting clear data layouts

Ziggy schemas help you define data layouts that are easy to grasp and validate for correctness. Ziggy schemas can define struct unions, enums and custom string literals.

The Ziggy LSP will be able to use your schema definitions to show diagnostics, contextual help and give autocomplete suggestions to your users.

Other tooling will be able to use schemas to autogenerate compatible type definitions.

The best feature of Ziggy schemas is that they can’t define every possible data layout!

package.ziggy-schema

$ = Package

struct Package {
  name: bytes,
  version: bytes,
  license: License,
  description: bytes,
  dependencies: {:}bytes,
  repository: Repository,

}

union License {
  /// Valid SPDX identifier 
  spdx: bytes,
  /// URL to license definition
  custom: bytes
} 

union Repository {
  git: Git,
  npm: Npm,

  struct Git { url: bytes, directory: ?bytes }
  struct Npm { name: bytes }
}

message.ziggy-schema

$ = Message

struct Message {
  id: Id,
  time: int,
  payload: Payload,

  union Id { uuid: bytes, seq: int }
  union Payload {
    command: Command,
    notification: Notification, 

    struct Command {
      do: Action,
      sender: bytes,
      roles: []bytes,
      /// Optional metadata. 
      extra: ?{:}bytes,
   
      union Action { clear_chat, ban_user: bytes }
    }

    struct Notification { text: bytes, lvl: int }
  }
}

Tooling Supremacy

One official CLI tool that solves the most common needs, like auto-formatting, schema validation, format conversion, querying (coming soon) and LSP support for comfortable editing in your favorite text editor or IDE.