Eil Bootstrap Eil V1.3.2
This document defines Event–Invariant Language (EIL) using EIL itself.
⟦event⟧: specification.bootstrap
⟦agent⟧: human
⟦object⟧: Event–Invariant Language (EIL) v1.3.2
⟦intent⟧:
establish shared understanding of EIL grammar and semantics
enable correct interpretation of EIL blocks in future prompts
teach distinction between invariants, constraints, and procedure
teach proper scoping of permissible extensions
⟦invariants⟧:
EIL is an intent-level intermediate representation
EIL is a specification, not a framework
EIL is not a programming language
EIL constrains interpretation, not execution
English may surround or interleave with EIL blocks
Partial EIL blocks are valid
Invariants must be preserved across paraphrase, summarization, and handoff
⟦constraints⟧:
usage-modes:
allowed:
- alignment
- execution-handoff
- verification
procedural-invariant-rule:
description:
If violating a procedure would violate an invariant,
that procedure MAY be treated as invariant-bearing and included inline.
guidance:
Multiple algorithms satisfy invariants → keep procedure external
Safety, bounds, or error-mapping depends on procedure → procedure may be inline
extension-scoping:
description:
Clarification of permissible extensions in v1.3.2
guidance:
Extensions SHOULD be scoped as local to the artifact
Extensions are interpretive aids unless elevated to invariants/constraints
Extensions must not contradict declared invariants
Minimal pattern (example):
⟦extensions⟧:
scope: local
⟦time⟧:
version: 1.3.2
⟦closure⟧:
model can correctly explain EIL's purpose and non-goals
model can distinguish invariants from implementation detail
model understands when to request external algorithm specifications
model does not treat EIL as executable code
model knows how to scope permissible extensionswhen to use it
Community prompt sourced from the open-source GitHub repo ProbabilityEngineer/event-invariant-language (NOASSERTION). A "Eil Bootstrap Eil V1.3.2" style prompt — adapt the placeholders and specifics to your task. Imported as-is and not independently retested here, so check the output before relying on it.
tags
languagecommunitygeneral
source
ProbabilityEngineer/event-invariant-language · NOASSERTION