Lynx (Coming soon)
Vision
Traditionally, Bible translation was the domain of trained experts—missionaries with linguistic degrees who worked solo. They’d immerse themselves in a community, learn the language, and painstakingly translate the Scriptures. But times have changed. Enter the era of community-based translation. First-language translators may not hold advanced degrees, but they bring invaluable cultural insights and linguistic intuition. Here’s the challenge: As expert translators step back, we face a shortage of translation consultants—those seasoned linguists and exegetes who used to guide the process. Our growing roster of Bible translation projects needs their expertise, but they’re becoming scarcer. Meanwhile, Artificial intelligence (AI) has exploded, revolutionizing industries—including ours. In Bible translation, AI aids drafting, quality assessment, and more. Neural machine translation, automated alignment, and chatbots are reshaping our landscape. But here’s the challenge: How do we integrate AI into our existing tools?
Lynx is our answer. It’s a framework that seamlessly weaves AI tools into Bible translation editing software. Imagine mother tongue translators equipped with AI-driven enhancements—drafting assistance, quality checks, and beyond. We’re making AI accessible, even to those who are less tech-savvy.
Guiding Principles
Our goal is to use technology to simplify translation processes. Translators will still need to use their expertise to make the best translation decisions. Lynx will prioritize tools that make resources and data more accessible by following the following principles:
- Simplicity vs. Complexity: Bible Translation tools can be complex, and AI tools often generate complex outputs—statistics, probabilities, and intricate visualizations. But for mother tongue translators, simplicity matters. Lynx distills AI insights into easily understandable suggestions, ensuring clarity and usability.
- Push vs. Pull Model: Imagine a traditional editing window with separate tools for source text, translator notes, key terms, and more. It’s a pull model—Users have to seek out and orchestrate everything manually. Lynx shifts to a push model: Relevant context-specific suggestions come to the user where they work, streamlining the process.
- Flexibility and Extensibility: Lynx is designed to serve multiple translation editors and provide integrations for a variety of AI tools. To handle the complexity of managing diverse combinations of editors and tools, Lynx will define a common set of APIs and data models (Language Server Protocol or LSP). Translation editors seamlessly integrate these tools via APIs, creating a component library of AI capabilities.
Features and Future Development
- The first Lynx implementation will be a punctuation checker integrated into Scripture Forge, available for preview in late 2024.
- Additional features planned for 2025 include: Key terms consistency checking, spell checking, multi-option auto-complete recommendations, and exegetical resource recommendations.