Skip to content
ES

More

ES

CASE STUDY · DOGFOODING

Nutrition Map × DARXI

A living experiment in project continuity.

EXPERIMENTAL
ONE RECORD, MULTIPLE RELATIONSILLUSTRATIVE EXAMPLE
Context
CTX-04
Evidence
EVID-A1
Decision
DEC-03
Unknown
UNK-07
Fictional identifiers. This example shows the record structure, not project data.
01The problem

A project accumulates more than code

Conversations. Decisions. Assumptions nobody wrote down. The reasons behind choices that now look arbitrary.

Most of it lives in places that do not survive: a chat window that closes, a thread nobody searches, a person who leaves.

TWO WAYS TO RESUME WORKCONCEPTUAL DIAGRAM

SCATTERED CONTEXT

Reconstruct the conversation.

A / ConversationB / EvidenceC / Context

WITH AN EXPLICIT RECORD

Follow the reasoning.

A / ConversationB / EvidenceC / Context
The record links sources, reasons and limits. It cannot recover what was never written.

Some of it does not resolve. That is recorded as an unknown.

02Journey

Five stages, with their real status

No stage is marked complete because it reads well. The last one remains undemonstrated.

STAGE 01

Context recovery

EXPERIMENTAL
GOAL
Produce a structured reading of what the project already knows.
LEARNING
A reconstruction is an interpretation, never a fact.
03Before / after

Nothing is added. It is arranged

BEFORE / AFTERSelect a relationship to compare

BEFORE

Knowledge in places that do not survive

AFTER

Knowledge in objects that persist

04What it preserves

Four objects, four different lifetimes

CONTEXT

Why decisions happened. The reasoning that is normally lost.

ILLUSTRATIVE EXAMPLECTX-04

EVIDENCE

What was observed, with its origin recorded and verifiable.

ILLUSTRATIVE EXAMPLEEVID-A1

DECISION

What humans chose. Including what they declined.

ILLUSTRATIVE EXAMPLEDEC-03

UNKNOWN

What remains open. Recorded as an object, not an absence.

ILLUSTRATIVE EXAMPLEUNK-07

05Decision lifecycle

A proposal is not a decision

Assistance proposes. A person reviews. What is decided stays — and what is rejected stays too.

TWO OUTCOMES · BOTH ARE PRESERVED
  1. 01

    Proposal

    Generated with citations to available evidence.

  2. 02

    Human review

    A person accepts, revises or rejects.

  3. 03

    Decision

    Human authority. Never automatic.

  4. 04

    Historical memory

    Kept with its reasoning, permanently.

APPROVED · WITH REASONREJECTED · WITH REASON

HUMAN DECISION

The record keeps the reasoning.

Approval does not turn evidence into absolute truth or execute an action by itself.

Rejected paths are kept too

Rejection does not erase the alternative.

A declined proposal stays in the record with its reason. Knowing what was not done, and why, is as useful as knowing what was.

06Boundary

Intelligence without secrets

DARXI preserves project intelligence without exposing project secrets.

PUBLIC

  • Evolution
  • Methodology
  • Learning
  • Results

PRIVATE

  • Source code
  • Business logic
  • Internal architecture
  • Sensitive information

The record shows that a decision was made, who made it, and what it rested on. It does not contain the system it was about.

This case study follows the same rule. What you are reading was written under a disclosure boundary.

07Signals

What is established and what is not

  • ESTABLISHED

    Decision lifecycle

    proposal → human ratification
    NOT ESTABLISHED

    External customer validation

    no external project absorbed
  • ESTABLISHED

    Human authority boundary

    structurally enforced
    NOT ESTABLISHED

    Long-term recovery

    not demonstrated over time
  • ESTABLISHED

    Evidence with provenance

    origin recorded and verifiable
    NOT ESTABLISHED

    Multi-team adoption

    single operator to date
  • ESTABLISHED

    Unknowns as objects

    recorded, never silently closed
    NOT ESTABLISHED

    Return on investment

    never measured
08Lab

This case is an experiment

Not a product demonstration. A notebook entry with its limits declared.

QUESTION
Can a project keep its context, decisions and unknowns in a form that survives the person who built it?
METHOD
Absorption of an internal project. Evidence, decisions and unknowns recorded as separate objects. No code changes to the project.
RESULT
The decision lifecycle and the authority boundary were demonstrated. Context recovery produced an interpretation, not a fact.
UNKNOWN
Whether this holds over time, with more people, and on projects that are not ours.
Lab →Evolution →
09Next

Bring a question

PARTICIPATE

If your organization loses decision context when tools or people change, you can propose a research question.

What “why” is disappearing on your team?

No delivery promise. No squad. One question can start a conversation.

[email protected] →See the lab →

YOUR NEXT QUESTION

Choose your depth.

Preparation

Explore possible usefulness, not success stories.

How it works

Understand the flow and what is preserved.

Laboratory

Examine the question, records and their limits.