Language Crosswalk
How does one concept stay stable across prose, shorthand, wire format, and diagram notation?
The Language Crosswalk is the binding layer. It keeps a DDL term, DML code, A&ID symbol, semantic layer, and operating use pointed at the same concept.
In the OKF meaning graph, the crosswalk protects compression. It lets agents move from glossary word to symbol to graph edge without losing what the idea means.
Binding Rule
Each crosswalk row should answer five questions:
- What is the readable term?
- What compressed code represents it?
- What diagram symbol represents it?
- Which semantic layer does it serve?
- What agent action changes because this binding exists?
Semantic Layers
Lexical:
- names what exists.
- example job: choose the right word.
Grammatical:
- defines how symbols combine.
- example job: parse a message.
Topological:
- places the concept in a system map.
- example job: trace dependencies.
Behavioral:
- changes an action or gate.
- example job: decide whether to proceed.
Intent:
- carries purpose, scope, or value.
- example job: decide whether action matters.
Checks
- The DDL term exists.
- The DML code has a decoder.
- The diagram symbol has one meaning.
- The semantic layer is useful.
- The binding changes retrieval, routing, or action.
Failure Modes
- Loose synonym — a row links similar words instead of one concept.
- Code without term — compressed notation cannot be explained.
- Symbol without action — a diagram mark decorates instead of guides.
- Layer blur — the row does not say whether it names, parses, maps, acts, or intends.
Context
- depends-on DDL Nomenclature — terms are named and defined there.
- depends-on Dreamineering Symbols — codes and symbols are decoded there.
- applies-to Open Knowledge Format Mapping — portable knowledge needs stable concept edges.
- pairs-with Platform — crosswalks become useful when systems can query them.
Questions
Which binding reduces translation cost most?
- Does the term exist?
- Does the code decode?
- Does the symbol point at one idea?
- Does the layer explain use?
- Does the binding improve agent action?