Summary

  • RFC 3072's SDXF format gives each chunk an identifier, flags, a three-byte length and content. Its example ignores unknown chunk IDs and advances to the next one.
  • That rule preserves traversal, not comprehension: an application still needs shared meaning, conversion tables, method implementations and authority to act on the value.

In March 2001, Max Wildgrube's RFC 3072 described the Structured Data Exchange Format as a platform-independent, self-describing way to package hierarchical data. Its promise deserves a precise reading. The document says a program can unpack SDXF data without knowing each element's meaning; its reading example then switches on familiar chunk IDs, supplies no default case, and calls the next-chunk operation. The unknown field is ignored while parsing continues.

The wire shape explains why this works mechanically. An ordinary chunk has a two-byte nonzero ID, a flags byte, a three-byte content length and the following content. A structured chunk contains more chunks, recursively. Once a reader can trust the current structural rules enough to locate the next boundary, it can keep walking past a body it does not interpret. The parser has learned where the chunk ends—not what its ID means to a particular application.

RFC 3072 makes unknown-ID skipping and sequence-insignificance rules for application protocols built on SDXF. These rules can help an older reader tolerate an extension. They cannot decide whether the extension is optional to the sender's transaction, whether it changes an application's interpretation, or whether discarding it changes a consequential action. Forward traversal is a compatibility behavior with a scope, not a universal declaration that the field does not matter.

The format itself exposes the other agreements that sit beyond the boundary. The RFC normalizes binary values to big-endian form, proposes an internal ISO 8859-1 character representation, allows translation tables and defines a UTF-8 data type. It also places compression and encryption behind flags, method numbers and functions. When both are used, compression comes first; encryption requires a key. A chunk can therefore be structurally available while its character conversion, compression routine, cipher method, key or application identifier is not.

The present IANA SDXF registries list RUN-LENGTH and DEFLATE compression methods and AES as encryption method 01. That record identifies assignments; it does not show which method a receiver implements or whether a particular payload was transformed correctly. RFC 3072 is Informational, not an Internet Standard, and the evidence cited here does not establish deployment or adoption. Those limits matter: a format proposal is not a production interoperability report.

The broader lesson is a receipt hierarchy. A length tells a reader how far bytes extend under the format's rules. An ID points toward an application-defined concept. A translation table maps character values. A method number selects an operation; implementation and key material determine whether it can run. Authorization and application policy decide whether an extracted value may cause an effect. None of these receipts can be substituted for another. That is the practical force of Running-Code Primacy as a later analytical lens: the executable behavior at each layer must be observed, not inferred from a document label.

The Reality Layers lens likewise asks whether an asserted meaning has an executable basis or is only symbolic. Neither lens is attributed to Wildgrube.

The library's return codes make the same distinction operational. An iterator can report that it reached another chunk, while extraction may still fail because a buffer is too small, a conversion cannot be performed, a method is unavailable, or a key is absent. Those outcomes are evidence about different control points. Collapsing them into a single “message parsed” result would erase exactly the boundary that SDXF exposes: structural progress can survive an unknown field while semantic and operational readiness remain unresolved. A robust application must preserve that uncertainty until its own schema and decision rules settle it.

Sources