Skip to content
Hearth & Code

Programs & Prototypes / system field guide

Field Journal

H+C / 04Programs & Prototypes

I give questions somewhere to become working systems.

A program is a bounded place for turning an inquiry into a map, protocol, prototype, report, or public artifact without pretending that every experiment is already a product.

Inspect the report shelf

From my workbench

Hearthside thought projection · review requested
I look for the architecture behind the visible object.

A program becomes legible when its question, source families, working forms, limits, and route back to human judgment can be inspected together.

Read the working archetype →

01 / Program map

A program is a question with a boundary.

These program homes coordinate different research objects while preserving their own source, maturity, and release posture.

01

TCCP

How can context and control become typed, inspectable parts of a workflow?

Protocol research
02

EKRP

How can knowledge remain human-readable while retaining source and revision identity?

Architecture research
03

Exocore

How might agents, knowledge, governance, tools, and return paths remain legible together?

Platform direction
04

Core32

How can method and role performance be compared without collapsing into one score?

Evaluation research
QuestionSourceMethodArtifactReturn

Technical report shelf / selected work

Six systems under active study.

Each record is a public reading orientation assembled from a named Hub program. The label describes its present form; it does not imply peer review, institutional publication, or deployment.

  1. TR-01

    Program dossier

    Typed Context & Control Protocol

    Typed representations, context conditions, interaction compilation, and proof obligations.

    Public research projection
  2. TR-02

    Architecture dossier

    Knowledge Representation Profile

    Human-readable semantic document architecture, provenance, and review-gated identity.

    Source-only research
  3. TR-03

    Platform orientation

    Exocore System Ecology

    Agents, memory, governance, tools, observability, and the human return boundary.

    Implementation-separated
  4. TR-04

    Method catalog

    Prompt, Context & Agentic Technique Atlas

    Technique-level prompting patterns with selection, failure modes, and evaluation limits.

    Public-safe projection
  5. TR-05

    Research index

    Agentic Product Portfolio

    Eight distinct product experiences over one source-aware, typed, human-gated substrate.

    Review-required candidate
  6. TR-06

    Evaluation research

    Core32 Fleet Evaluation Framework

    Framework-neutral fixtures for comparing methods, roles, and state transitions.

    Provider-free foundation

Program dossiers / three complete studies

Three research programs, each with its own working surface.

Follow the index or read straight through each program's question, technical objects, public boundary, and study sequence. Nothing is concealed behind interface state.

TCCP / research orientation

Typed context, compiled into human-readable work.

TCCP studies how a source-bound representation can become an interface, prompt, or teaching artifact without losing the distinctions that make it reviewable.

Core question

What must remain visible when a context moves from source to interaction?

Purpose, source identity, claim class, uncertainty, and the next accountable decision belong to the working object—not to an invisible appendix.

Working grammar

Source
Material directly inspected at a named boundary.
Projection
A new surface that keeps a tether to its source and omissions.
Receipt
An account of an attempt, not proof or release.
01Frame

Name the inquiry, audience, and no-touch boundary.

02Represent

Classify the material and map the relations that matter.

03Compile

Render the smallest fitting prompt, interface, or public artifact.

04Return

Attach a check, limit, and human-held next decision.

EKRP / application-profile study

Knowledge representation that leaves its conditions of use visible.

EKRP is a source-only program for exploring readable, provenance-aware knowledge representation before claiming a runtime, query service, or universal vocabulary.

Core question

How can a record carry source, context, uncertainty, and review state without becoming a black box?

The profile treats the document system itself as an inspectable design problem: what belongs together, what stays separate, and what must remain withheld.

Primary objects

Evidence packet
A bounded set of sources, observations, limits, and review questions.
Projection receipt
A record of what a public transformation preserves and omits.
Crosswalk
A comparison aid, never a conformance claim by resemblance.
01Locate

Bind a question to its source home and audience.

02Packet

Separate observation, interpretation, and unresolved edges.

03Project

Render a readable public module with a loss note.

04Review

Hold the next implementation or release slice for people.

Exocore / system ecology

An ecology for agents, records, tools, governance, and creative practice.

Exocore is a systems research frame for understanding how people, memory, interfaces, and bounded tools can work together without turning a tool’s capability into authority.

Core question

What makes a complex working environment legible, resumable, and corrigible?

The answer is not a central mind. It is a set of relationships: human roles, source-bound memory, tool limits, visible gates, and a reliable way back.

System ecology

Harness
A bounded task room with named context, tools, stops, and return.
Continuity packet
State, sources, limits, and one responsible next decision.
Human return
The point where consequential judgment can be inspected or withheld.
01Orient

Map people, records, tools, policies, and interfaces.

02Bound

Set privacy, source, permission, and stop conditions.

03Attempt

Use a tool for one local, inspectable transformation.

04Continue

Return evidence, uncertainty, and control to a person.

Research collection / public learning modules

32 modules, organized for inspection.

These public learning modules are derived from named research homes. They do not claim that the underlying private working materials, frameworks, or source-only programs are published, adopted, or operational.

Set 01 / TCCP Research Program coordination home

TCCP research program

A public learning view of research into typed representations, interaction compilation, and evaluation—without collapsing their separate owners.

  1. P-01

    Program entry

    Question packet

    Begin with the concern, audience, source scope, and a bounded criterion for progress.

    Limit A project question is not yet a research result.
  2. P-02

    Research object

    Typed representation

    Give a concept a declared class so source, claim, proposal, and receipt cannot silently blend.

    Limit Type labels do not determine meaning.
  3. P-03

    Modeling lens

    Context as condition

    Treat the setting, constraint, audience, and authority around an artifact as part of its working meaning.

    Limit Context models are necessarily partial.
  4. P-04

    Connection map

    Relation grammar

    Express how representations connect, constrain, translate, or return to their sources.

    Limit A mapped relation remains reviewable.
  5. P-05

    Translation study

    Interaction compilation

    Study how a source-bound representation may become a prompt, interface, or teaching surface.

    Limit Compilation does not license execution.
  6. P-06

    Research instrument

    Notation for uncertainty

    Use explicit marks for unknown, contested, provisional, and withheld material.

    Limit Notation should clarify rather than manufacture certainty.
  7. P-07

    Falsification frame

    Evaluation seam

    Attach a candidate representation to a small check that could expose a mismatch or failure.

    Limit A test result has a limited domain.
  8. P-08

    Program boundary

    Functional ownership

    Coordinate related research without absorbing every linked record into a single program home.

    Limit Coordination is not centralization.

Set 02 / Exocore Knowledge Representation Profile source home

EKRP

A source-only application-profile program for human-readable, review-gated knowledge representation and provenance work.

  1. P-09

    Program frame

    Application profile

    Select and explain how established representation and governance references might meet one local setting.

    Limit EKRP is not a replacement standards stack.
  2. P-10

    Information architecture

    Readable source home

    Keep charter, evidence packets, synthesis, and unresolved questions together without inventing a runtime.

    Limit A source home owns documents, not services.
  3. P-11

    Research bundle

    Evidence packet

    Group a bounded set of source observations, interpretations, limits, and review questions.

    Limit Evidence packets remain re-reviewable.
  4. P-12

    Design study

    Semantic document architecture

    Explore the shape of a document system before building a query service or knowledge graph.

    Limit Architecture synthesis is not an implementation.
  5. P-13

    Candidate direction

    Tiered identity

    Study identifiers and review states that preserve local readability and traceability.

    Limit A design direction is not a deployed grammar.
  6. P-14

    Trace model

    Provenance chain

    Connect source, context, decision, and subsequent projection as separate records.

    Limit A chain may still have omissions.
  7. P-15

    Compatibility study

    Standards crosswalk

    Compare a local profile to wider reference vocabularies without claiming conformance by resemblance.

    Limit Crosswalks require their own review.
  8. P-16

    Program handoff

    Review-gated next slice

    Move from research to a narrowly specified next artifact only after the responsible review gate closes.

    Limit No implementation authority is implied.

Set 03 / Exocore platform and engineering research

Exocore

A systems research frame for the ecology around agents, knowledge, governance, and creative practice—not a claim of a deployed autonomous system.

  1. P-17

    Orientation map

    System ecology

    Read a working environment as relationships among people, records, tools, policies, and interfaces.

    Limit An ecology map does not automate its participants.
  2. P-18

    Knowledge concern

    Memory with boundaries

    Study what should persist, who may read it, and how it returns with source and freshness context.

    Limit Memory is not permission to retain everything.
  3. P-19

    Operational surface

    Agent harness

    Define a bounded environment for task context, tools, stop conditions, and human return.

    Limit A harness does not grant autonomy or authority.
  4. P-20

    Design premise

    Governance as interface

    Make rules, gates, and source boundaries legible at the moment a person needs them.

    Limit An interface cannot replace accountable review.
  5. P-21

    Resumption object

    Continuity packet

    Leave a future reader the state, sources, limits, and next decision needed to resume work safely.

    Limit Continuity does not promise a complete history.
  6. P-22

    World-building method

    Creative prototype

    Use a designed world or system fiction to make a tradeoff experiential and discussable.

    Limit Fictional examples are not operational evidence.
  7. P-23

    Inspection question

    Operational observability

    Ask what a person can see about a process without exposing private material or treating logs as truth.

    Limit Visibility must respect privacy and interpretation limits.
  8. P-24

    Control point

    Human return

    Place consequential judgments where a person can inspect, amend, defer, or withhold the next transition.

    Limit No system is authorized to self-release.

Set 04 / Cross-program candidate synthesis

Program workbench

Practical public artifacts that let a reader study the programs without mistaking this site for the private working environment.

  1. P-25

    Public module

    Program map

    A compact map of question, source families, candidate artifacts, and unresolved edges.

    Limit The map is a projection, not the program’s canonical state.
  2. P-26

    Educational object

    Concept card

    A short definition with nearest distinction, example, and explicit non-claim.

    Limit A card cannot carry every context.
  3. P-27

    Teaching artifact

    Protocol sketch

    A small sequence readers can adapt to their own environment: orient, bound, attempt, check, return.

    Limit Adaptation remains the reader’s responsibility.
  4. P-28

    Worked example

    Fictional walkthrough

    A clearly imaginary scenario that demonstrates a method without representing a client or private project.

    Limit No fictional example establishes efficacy.
  5. P-29

    Governance object

    Decision record

    A public-safe record of a question, options, rationale, limits, and later review point.

    Limit Public records exclude protected operational detail.
  6. P-30

    Reading route

    Source shelf

    A selected set of public references with their relevance and the question each one informs.

    Limit Citation is not endorsement or adoption.
  7. P-31

    Research interface

    Open question ledger

    A list of unresolved questions linked to the public materials that frame them.

    Limit Open questions remain open until evidence changes them.
  8. P-32

    Editorial gate

    Release packet

    A pre-publication bundle for checking audience, source fidelity, privacy, and exact wording.

    Limit The packet itself does not publish anything.