Hello World in SPARK Ada - Ada-spark Typing CST Test

Primary Use Cases

• ▸Avionics flight control software
• ▸Railway signaling and interlocking
• ▸Medical device firmware
• ▸Cybersecure embedded systems
• ▸Automotive safety ECUs

Notable Features

• ▸Design-by-contract with pre/postconditions
• ▸Strong static typing and modularization
• ▸GNATprove formal verification
• ▸Guaranteed absence of runtime errors (AoRTE)
• ▸High-integrity code generation support

Origin & Creator

Developed by Praxis (now Altran UK) and later extended by AdaCore, leading vendors of Ada technology for mission- and safety-critical software.

Industrial Note

SPARK is used where DO-178C Level A or IEC 61508 SIL-4 compliance is required, enabling mathematical proof of correctness for safety-critical and security-critical applications.

Quick Explain

• ▸SPARK is a subset of Ada that removes unsafe or ambiguous features.
• ▸Uses contracts (pre/postconditions, invariants) for formal correctness.
• ▸GNATprove tool performs static analysis and proofs.
• ▸Designed for avionics, medical, defense, and railway systems.
• ▸Guarantees freedom from runtime errors (overflow, null dereference, etc.) through proofs.

Core Features

• ▸Packages, procedures, functions
• ▸Subtype constraints and range checks
• ▸Contracts (pragma Assume, Assert, Post, Pre, Loop_Invariant)
• ▸SPARK proof annotations
• ▸Pure, deterministic code model

Learning Path

• ▸Learn Ada fundamentals
• ▸Understand SPARK restrictions
• ▸Add simple contracts
• ▸Use GNATprove
• ▸Master formal program design

Practical Examples

• ▸Prove array bounds safety
• ▸Build a verified stack ADT
• ▸Prove correctness of a PID controller
• ▸Motor controller with no runtime errors
• ▸Train braking logic verification

Comparisons

• ▸SPARK vs Ada: SPARK is the provable safe subset.
• ▸SPARK vs Rust: Rust focuses on memory safety; SPARK proves full correctness.
• ▸SPARK vs MISRA C: SPARK is more rigorous and verifiable.
• ▸SPARK vs SCADE: SPARK is code-level proofs; SCADE is model-based.
• ▸SPARK vs Eiffel: SPARK contracts are verifiable, not runtime only.

Strengths

• ▸Mathematical proof of correctness
• ▸Ultra-reliable for safety-critical domains
• ▸Eliminates runtime exceptions
• ▸Highly readable and maintainable syntax
• ▸Standardized and long-supported language

Limitations

• ▸Restricted subset compared to full Ada
• ▸Verification requires discipline and learning
• ▸Not suitable for rapid UI/desktop apps
• ▸Runtime features like dynamic allocation limited
• ▸Build times grow with extensive proofs

When NOT to Use

• ▸Rapid prototyping or scripting
• ▸Feature-rich GUIs or web apps
• ▸Highly dynamic memory systems
• ▸Non-critical hobby projects
• ▸Where verification costs outweigh benefits

Cheat Sheet

• ▸pragma SPARK_Mode(On)
• ▸with Pre => Condition
• ▸with Post => Condition
• ▸Loop_Invariant => Expression
• ▸Run: gnatprove -P project.gpr

FAQ

• ▸Is SPARK a separate language? -> No, a subset of Ada.
• ▸Can SPARK prove all logic? -> No, but proves safety properties.
• ▸Does SPARK allow pointers? -> Strictly controlled, mostly disallowed.
• ▸Can you mix Ada and SPARK? -> Yes, with boundaries.
• ▸Is runtime checking needed? -> Often eliminated after proofs.

30-Day Skill Plan

• ▸Week 1: Ada basics
• ▸Week 2: SPARK subset rules
• ▸Week 3: Contracts and subtypes
• ▸Week 4: GNATprove practice
• ▸Week 5: Verified ADTs and loops

Final Summary

• ▸SPARK is a verifiable subset of Ada for ultra-safe systems.
• ▸Used in avionics, defense, rail, and medical industries.
• ▸GNATprove enables mathematical correctness proofs.
• ▸Eliminates runtime exceptions and logic flaws.
• ▸Ideal where safety and reliability are non-negotiable.

Project Structure

• ▸src/ for code
• ▸gpr project file
• ▸spec (.ads) and body (.adb) files
• ▸proof annotations
• ▸test and harness modules

Monetization

• ▸Safety-critical software consulting
• ▸DO-178C verification services
• ▸Railway or avionics tool integration
• ▸Formal verification training
• ▸Certified code development

Productivity Tips

• ▸Write contracts early
• ▸Use subtypes to avoid runtime errors
• ▸Run proofs continuously
• ▸Minimize global dependencies
• ▸Keep procedures small and clear

Basic Concepts

• ▸Packages - modular code units
• ▸Subtypes - type-safe constraints
• ▸Contracts - correctness rules
• ▸Proof obligations - automatically generated checks
• ▸Abstract data types - encapsulation for safety

Official Docs

• ▸https://www.adacore.com/sparkpro
• ▸https://learn.adacore.com/courses/SPARK_for_the_MISRA_C_Developer