Specification reference

English · 简体中文

These documents define object formats and composition semantics. When implementing or validating the underlying tools, treat the rules here and the corresponding tests as authoritative.

• Language and algorithm library conventions
  • The Python generation layer
  • Operations and modules
  • Algorithm protocols
  • Matrix representation and normalization
  • Data access
  • Call capabilities
  • Precision and readout
  • Serialization and backends
  • Versions and compatibility
• RIR: the modular register-level intermediate representation
  • Part 1: Design positioning
  • Part 2: Data model
  • Part 3: Instruction set
  • Part 4: Program-level rules
  • Part 5: Complete examples
• QRAM memory definition (qram YAML)
  • Fields
  • Semantic rules
  • Python API
  • CLI
  • Schema
• Open declarations, capabilities, and binding
  • An open declaration is not an identity operation
  • Capabilities and isometric state preparation
  • Binding
  • QRAM capture
  • The boundary of openness
  • Migration from the old rules
• MIR 0.1: the pure mathematical function graph
  • Objects and types
  • Finite numerics and generation configuration
  • Outputs and status
  • Front-end boundary
• General QHAM: derivation from PDEs to quantum-adapted linear systems
  • 1. Input scope and conventions
  • 2. Deriving the higher-order recursion from the homotopy equation
  • 3. Tensor variables on independent coordinates
  • 4. The finite closure proof
  • 5. The physical solution and forcing homogenization
  • 6. Validation method and boundaries
• Replacing the QLSS inside QFVM: a review of input models, QRAM data structures, and output contracts
  • Conclusions
  • 1. What the three papers actually assume
  • 2. The four layers that must be kept apart
  • 3. Baseline problems and this round’s fixes
  • 4. How the new QFVM sparse input is implemented
  • 5. The sparse-to-BE adaptation is not a type conversion
  • 6. Output replacement also needs a contract
  • 7. The actual scope of QRAM and quantum data structures
  • 8. How to replace
  • 9. This round’s evidence and what remains unfinished
  • 10. Numerical validation record (2026-09-16)
• OriginIR-ext and PySparQ review
  • OriginIR-ext strengths and boundaries
  • PySparQ’s register model
  • Compatibility validation in the recorded validations
  • 2026-09-09 addendum: application-description acceptance
  • 2026-09-09 custom-operator path