Roadmap: build RockQL from compiler MVP to v1.0

[SayanthRock]

Goal

Build RockQL into a readable data-pipeline language, compiler, local database playground, visual editor, and cross-platform developer tool without expanding into database administration or write operations.

Compiler foundation starts in #1.

v0.1 — Compiler foundation

• Rust workspace and crate boundaries.
• AST with source spans.
• Parse from, filter, select, derive, sort, and take.
• Multiline and pipe-separated source.
• Generic SQL, SQLite, and PostgreSQL output foundation.
• CLI commands: compile, check, ast, and format.
• Formatting, Clippy, tests, release build, and example CI.
• Add a real expression lexer and parser instead of preserving expression text.
• Add identifier, literal, operator, function-call, date, range, and null-coalescing AST nodes.
• Add resolver diagnostics for unknown names and invalid transformation order.
• Add relational intermediate representation.
• Add snapshot test harness and dialect fixtures.
• Add MySQL and DuckDB generators.
• Add signed Linux, Windows, and macOS release binaries with checksums.

v0.2 — Web playground

• WebAssembly compiler package.
• React, TypeScript, and Vite application.
• Monaco RockQL syntax highlighting.
• RockQL and generated SQL side by side.
• Inline diagnostics with exact spans.
• Dialect selector.
• Formatter and copy/export actions.
• CSV and JSON import.
• Local DuckDB-Wasm execution.
• Results, schema, explain, errors, and history panels.
• Shareable query links.
• Dark, light, mobile, tablet, and desktop layouts.

v0.3 — Data operations

• join.
• group and aggregate.
• distinct, union, skip, and let.
• Aggregate functions: count, sum, average, minimum, and maximum.
• Variables, function calls, dates, ranges, and ??.
• Name and type resolver.
• PostgreSQL, MySQL, and DuckDB conformance tests.

v0.4 — Developer tooling

• Dedicated formatter crate.
• Language Server Protocol implementation.
• VS Code extension.
• Tree-sitter grammar.
• Completion, hover, diagnostics, formatting, and go-to-definition.
• JavaScript, Python, Java, and C binding foundations.

v0.5 — RockQL Studio

• Visual pipeline editor.
• Code-to-block and block-to-code synchronisation.
• Source, filter, select, derive, join, group, aggregate, sort, limit, union, and output blocks.
• Schema browser.
• Local DuckDB execution.
• CSV, JSON, and Parquet support.
• Query history and reusable snippets.
• Reviewable AI diffs for query generation, explanation, fixing, and optimisation.

v1.0 — Stable language and applications

• Stable syntax specification.
• Compatibility and deprecation policy.
• Full reference documentation and book.
• Compiler benchmarks and performance budgets.
• Signed cross-platform releases.
• Android app using Kotlin, Compose, Material 3, Room, and DataStore.
• Android SQLite/SQLDelight execution and local file import.
• Phone, tablet, and foldable layouts.
• Biometric protection for saved credentials.
• Desktop application.
• Stable JavaScript and Python bindings.

Required workflows as components land

• ci.yml.
• security.yml.
• rust-tests.yml.
• wasm-build.yml.
• web-build.yml.
• android-build.yml.
• docs.yml.
• release.yml.

Boundaries

Do not add database administration, users/roles, migrations, INSERT, UPDATE, DELETE, cloud credential storage, team collaboration, paid AI services, or complex optimisation during the early releases.

Third-party code must retain its original licence, copyright notices, and attribution. RockQL branding, syntax extensions, compiler implementation, and user experience must remain distinct.