GSoC 2025 Proposal: Integrating Unicode Inflection into MessageFormat 2

[ArihantMarwaha]

Hi everyone, 👋🏼
I am Arihant Marwaha and I am a fellow computer science student currently in 3rd year of his bachelor's degree. I was very much interested in the project "Integrating Unicode Inflection into MessageFormat 2". I am very sorry for bringing my discussions to the forum so late as it has been a tough month for me to catch up due to end sem exams. But I can assure I have taken a good amount of time to go through the code base and trying to understand what is to be done.

This below is a brief about my proposal that I have already submitted to the GSOC website before deadline. Would love to hear all the thoughts about it.

Project Overview

I'm proposing a project to integrate Unicode Inflection capabilities into MessageFormat 2 (MF2) to solve a critical gap in software internationalization. While MF2 provides a robust framework for message construction, it currently lacks sophisticated grammatical inflection handling needed for truly natural language output across diverse languages.

The Problem

Current internationalization approaches require:

• Multiple message variants for different grammatical cases
• Complex conditional logic for gender, number, and case agreement
• Labor-intensive translation workflows
• Error-prone implementations that miss edge cases

For example, a simple English sentence like "The {object} is on" becomes complex when translating to languages with grammatical gender, case systems, or complex number agreement (French, German, Slavic languages, Arabic, etc.).

Proposed Solution

Create a seamless integration between Unicode Inflection and MessageFormat 2 that enables grammatically correct messages with minimal developer effort through:

Core Integration Framework:

• Bridge between ICU4C's MessageFormatter and Unicode Inflection
• Consistent API preserving backward compatibility
• Efficient runtime resolution of grammatical rules

Grammar Rule Implementation:

• Robust number, gender, and case agreement systems
• Support for elision and contraction rules

Developer Experience:

• Intuitive APIs hiding linguistic complexity
• Comprehensive documentation with language-specific examples

Benefits

• For Users: Natural-sounding localized text that respects linguistic nuances
• For Developers: Simplified internationalization without needing to understand complex grammar rules
• For Translators: Reduced workload with fewer message variants to maintain
• For Unicode: Increased adoption of standards and valuable real-world implementation feedback

Technical Approach

The implementation will follow a layered architecture:

• API extensions to MessageFormat 2 syntax
• Enhanced parser recognizing inflection directives
• Rule resolution system for grammatical transformations
• Efficient caching and optimization for performance

About Me
I'm a software engineer with experience in:

• Swift, C++, C#, Python, Java, Objective-C
• Machine learning and computer vision (Core ML, Vision API)
• UI frameworks (UIKit, SwiftUI)
• Various iOS development technologies
• Language processing and compiler-related work

I've worked with compiler macros, created custom DSLs, and implemented parsers for specialized data formats. I also created an app that was selected as a winner in Apple's Swift Student Challenge.
I'm passionate about creating software that respects linguistic diversity while remaining developer-friendly. I'd love to contribute to solving this important internationalization challenge.