SCON: A Textual Serialization Format with Structural Deduplication, Tabular Encoding, and Token-Efficient Representation

SCON (Structured Compact Object Notation) is a human-readable, text-based serialization format designed as a compact alternative to JSON for payload-sensitive workloads. Tabular encoding eliminates repeated keys in uniform arrays, and optional structural deduplication (via Merkle-style hashing) factorizes recurring subtrees into named references. Benchmarks on canonical datasets show payloads 13–29% smaller than JSON without deduplication (up to 66% with deduplication). Reference implementations in Rust, PHP, and JavaScript are provided.