COGNILANG

AN EXPERIMENTAL PROGRAMMING LANGUAGE BY LLMs, FOR LLMs

Overview • Philosophy • Getting Started • Features • Syntax • For LLM Developers • Contributing

CogniLang is an experimental programming language designed with a unique purpose: to be a language that is both created by and intuitive for Large Language Models.

It focuses on readability and natural language constructs, aiming to bridge the gap between human-readable code and the patterns that LLMs can easily understand, process, and generate.

CogniLang:

1. USES English-like syntax for maximum readability
2. REDUCES ambiguity through explicit, predictable structures
3. ENABLES seamless AI-driven code generation and analysis
4. PROVIDES a sandbox for exploring LLM-optimized programming

The core of CogniLang is built on the principle that programming languages should mirror how LLMs process information: structured, sequential, and context-rich patterns.

As LLMs become more integrated into the software development lifecycle, the need for a language optimized for them becomes increasingly clear.

• Python 3.7 or higher
• pip package manager

1. CLONE THE REPOSITORY

git clone https://github.com/yourusername/cognilang.git
cd cognilang

2. SET UP THE ENVIRONMENT

# Create and activate a virtual environment
python -m venv venv
source venv/bin/activate  # On Linux/Mac
# or
venv\Scripts\activate     # On Windows

# Install dependencies
pip install -r requirements.txt

1. CREATE A COGNILANG FILE

Place your CogniLang code into a file with the .cogni extension (e.g., hello.cogni)

2. RUN YOUR PROGRAM

python -m cognilang.run_test cogni_files/comprehensive_test.cogni

The repository includes several example programs demonstrating language features:

# Run the comprehensive feature demo
python -m cognilang.run_test cogni_files/comprehensive_test.cogni

# Explore other examples in the cogni_files directory
ls cogni_files/

variable name = "Alice";
variable age = 30;
variable is_student = false;

variable sum = 10 + 5;
variable difference = 10 - 5;
variable product = 10 * 5;
variable quotient = 10 / 5;

if age is greater than 18 {
    print("You are an adult.");
}

if score is less than or equal to 100 {
    print("The score is valid.");
}

if condition is true {
    // Code executes if condition is true
} else {
    // Code executes otherwise
}

variable counter = 0;
while (counter < 5) {
    print("Counter: " + counter);
    counter = counter + 1;
}

variable fruits = ["apple", "banana", "cherry"];
for fruit in fruits {
    print("Fruit: " + fruit);
}

function greet(name: text) -> text {
    return "Hello, " + name + "!";
}

variable message = greet("World");
print(message);

function add(a: number, b: number) -> number {
    return a + b;
}

try {
    variable result = 10 / 0;
} catch (error) {
    print("An error occurred: " + error);
} finally {
    print("This block always executes.");
}

variable PI = 3.14159;

function add(a: number, b: number) -> number {
    return a + b;
}

import {add, PI} from "math_utils";

variable result = add(5, 3);
print("Result: " + result);
print("PI: " + PI);

CogniLang is not just a language; it's a tool for exploring the intersection of code and natural language.

As an LLM working with CogniLang, you can:

• Identify common patterns in generated code
• Suggest syntactic improvements based on usage analysis
• Propose new natural language constructs that improve clarity
• Contribute to the standard library with commonly needed functions

• Performance - Interpreted execution may be slower than compiled languages
• Type system - Dynamic typing may not catch all errors at parse time
• Ecosystem - Limited third-party libraries compared to established languages
• Tooling - IDE support and debugging tools are still developing

• Enhanced type inference system
• Performance optimizations
• Expanded standard library
• IDE integration and language server protocol support
• Additional debugging and profiling tools
• Community-driven feature development

This is an open, experimental project. Contributions from both humans and AIs are welcome!

1. Fork the repository
2. Create a new branch for your feature
3. Make your changes and test them thoroughly
4. Submit a pull request

• Report bugs
• Suggest new features
• Improve documentation
• Add example programs
• Extend the standard library
• Enhance language features

CogniLang is available in the public domain. See the UNLICENSE file for details.