Sarvalekha / Shared product foundation

Sarvalekha Foundation

In development

Sarvalekha Foundation is the common development framework for the Sarvalekha ecosystem. Its independently usable Core brings schema-defined models, generated codecs and selected managed state, cache and embedded storage capabilities together. The renamed 1.0.0 framework remains in development.

Under development — not released. No supported release or subscription is established.

Foundation product emblem

Implemented development capabilities

A foundation to explore.

Development scope · not released

Start with a shared model

Generate models and codecs for C++, C, Java, C#, JavaScript, TypeScript, Go, Rust, Python, Swift and Kotlin. JSON and native binary formats have documented profiles; selected C++ outputs add Protobuf-related formats. Backend and compression support varies.

Development scope · not released

Add managed application state

The C++ managed runtime supplies typed editing, transactions, immutable reads, linear or branching histories and selected durable journals. Other language outputs have narrower model, platform and recovery boundaries.

Development scope · not released

Coordinate with explicit boundaries

Selected authority/client mechanisms coordinate accepted edits, retries and merges. Applications supply authentication, trusted sessions, transport, scheduling and conflict resolution; cross-service ACID is outside this scope.

Development scope · not released

Work with structured document formats

Selected profiles provide XML mappings, XSD operations and ZIP/OPC typed document parts with bounded opaque content. These file-format facilities do not provide a document renderer.

Development scope · not released

Choose the right storage and reuse profile

Optional C++ storage persists one managed root with an exclusive writer and immutable snapshots. Scalar and page-backed indexes, rebuildable values and data caches retain explicit validity, ownership and resource limits.

Know the boundaries

Scope matters.

  • Under development, with no published Foundation SDK artifact established by the source rename. Compiler/runtime, schema language and paired editor identities begin at 1.0.0; broad qualification and paired release remain separate.
  • Core can be used independently for data exchange without a product engine, cloud credentials or running service. Applications retain domain logic, user interfaces, authentication, transport and service authority.
  • Native components require C++23. Language and provider profiles have distinct prerequisites and do not establish universal parity.
  • The clean Foundation contract has no legacy adapters, old-data converter or automatic artifact-pin upgrade. Consuming applications require matched dependencies and separate qualification.
  • Eager storage retains complete roots. Page-backed indexes do not establish larger-than-RAM editing or a process-wide memory ceiling; no general SQL, graph, spatial or vector engine is implied.

A place in the wider picture

Connected, with clear roles.

What comes next

Direction, without a deadline.

  • Qualify the renamed Windows/Linux framework and paired editor release
  • Optional Database, Platform and Diagnostics modules with common tooling
  • Broader language backends, protected caches, storage and service adapters with separate implementation and qualification
View the product roadmap