CONFLANG 2026

CONFLANG is a workshop dedicated to the design, theory, practice, and future evolution of configuration languages.

Modern software systems, including build systems, package managers, operating systems, cloud infrastructures, and web services, require increasingly complex and non-trivial configuration. This complexity is often shifted from code to configuration artifacts, making configuration a critical locus for concerns such as security, reliability, reproducibility, and maintainability. The rise of declarative paradigms, such as infrastructure-as-code, has further amplified the importance of configuration as a first-class software engineering concern.

However, traditional static, text-based configuration formats remain fundamentally limited. They lack abstraction mechanisms, composability, and principled validation. As a result, developers frequently resort to ad hoc solutions, leading to duplication, inconsistencies, and fragile pipelines where errors are detected too late if at all.

To address these limitations, a new generation of configuration languages has emerged. These range from programmable configuration systems that integrate computation and abstraction, to specialized languages and tools focused on validation, synthesis, or transformation. These languages occupy a distinct and underexplored region of the design space, shaped by unique constraints such as reproducibility, auditability, partial evaluation, and integration with heterogeneous systems.

At the same time, recent advances in AI and large language models (LLMs) are beginning to reshape how configuration is authored, analyzed, and maintained. LLMs are increasingly used to generate configuration artifacts, assist in migration between formats, detect misconfigurations, and even synthesize repair actions. This raises new research questions at the intersection of programming languages, software engineering, and AI:

• How can configuration languages be designed to be robust under AI-assisted generation?
• How can we ensure correctness, safety, and explainability of AI-generated configurations?
• What abstractions are needed to integrate synthesis, validation, and repair into configuration workflows?
• How can configuration languages serve as an interface between human intent and machine-generated artifacts?

CONFLANG aims to bring together researchers and practitioners working on these challenges. The workshop fosters exchange between academia and industry, providing a venue to discuss novel ideas, share experiences, and identify emerging research directions. Configuration remains a pervasive and industrially critical problem domain, and we believe it offers fertile ground for impactful and interdisciplinary research.

Call for Contributions

CONFLANG 2026 welcomes a broad range of contributions on the design, implementation, analysis, and use of configuration languages and related systems.

We explicitly aim to bridge programming languages, software engineering, DevOps practices, and AI-assisted development, and encourage both academic and industrial perspectives.

Topics of Interest

Topics include, but are not limited to:

• Design and semantics of configuration languages
• Programmable and declarative configuration systems
• Validation, verification, and static analysis of configurations
• Configuration testing and debugging
• Configuration synthesis and transformation
• Managing variability and reuse in configurations
• Configuration for cloud, infrastructure, and distributed systems
• Security and compliance in configuration
• Empirical studies and experience reports
• Tooling and developer experience

Potential AI-related topics:

• LLM-based generation of configuration artifacts
• AI-assisted configuration repair and validation
• Prompting and interaction models for configuration tasks
• Ensuring correctness, safety, and robustness of generated configurations
• Benchmarks and datasets for AI in configuration
• Hybrid approaches combining formal methods and AI Submission Types

We invite two types of contributions:

1. Extended Abstracts (Non-archival)

   • Up to 2 pages (≈600–1000 words)
   • Not included in formal proceedings
   • Reviewed lightly by the chairs for relevance and quality
   • Suitable for: Talks and presentations, Experience reports, Position papers

2. Papers (Archival, Proceedings)

   • Short papers: up to 4 pages
   • Long papers: up to 8 pages
   • Included in the workshop proceedings
   • Must present original, unpublished work