Manifesto for Collaborative Concurrent Extreme Software and Product Development for humans
This document is fully meant for human readers; it is not meant for robots, automation pipelines, agents, or chatbots. This is a set of observed practices and philosophy within IT development.
Digital knowledge grows through materialisation of situations and events.
A solution can be copied and altered without being consumed.
An idea can be improved without being diminished.
The organization is designed to remain coherent while many participants continuously create new situations in the same shared environment.
Principles
Understanding
Understanding is the foundation of capability.
A system is understood when people can spend time talking about production matters in human language.
Knowledge must be distributed throughout the organisation between peers and in common channels.
Understanding grows through participation, observation, repair, and continuation.
Contribution
Everyone can contribute.
Initiative creates situations.
Delivered solutions create situations.
Responsibility follows participation and involvement.
Contribution begins before completion.
A note, commit, issue, demo, test, sketch, failure, or unfinished branch can become useful material at any moment and within new situations.
Visibility
Problems, discoveries, decisions, experiments, failures, and repairs should remain visible.
Shared information may lead to shared understanding.
Shared understanding creates collective capability.
Machines may clarify and compress work, but they must not hide the effort, uncertainty, judgement, sacrifice, and recovery behind it.
Concurrency
Work should proceed concurrently.
People may explore and modify the same area at the same time.
Overlap creates information. Integrity is maintained through communication, visibility, and participation.
Differences reveal assumptions. Multiple explorations reveal possibilities faster than a single path.
Conflicts reveal dependencies. Experiments transform disagreement into learning.
The goal is maintenance of integrity while change is occurring and visible participation in interference.
The measure of a system is its ability to evolve while many people change it simultaneously in continuous iterations.
Communication
Code, documentation, interfaces, conversations, demos, and artifacts are tools for transferring understanding.
Clear communication increases the ability to collaborate and evolve systems.
Every artifact should suggest a next move.
Work should be readable, testable, challengeable, reusable, and transformable.
Workshop
Work belongs to a shared bench, where individual authorship matters.
The form must survive touch by others — make things that can be entered, questioned, extended, repaired, and continued.
Art is welcome, private fog is not.
Strange work should invite deeper reading rather than protect weak thinking.
Learning
Feedback drives change.
Every release, incident, success, failure, experiment, and repair increases available information.
Learning compounds over time.
Speed accelerates production.
Slowness enables comprehension.
Use machines to accelerate formatting and exploration while preserving the human pace required for understanding.
Stewardship
Leadership emerges through contribution and understanding.
Management connects people, coordinates effort, exposes dependencies, and reduces friction.
Anyone able to change a system should be able to participate in its stewardship.
Ownership is not possession.
Care includes allowing others to use, repair, question, remix, and continue the work.
Collective Juggling
Nobody holds everything.
Capability emerges from timing, attention, awareness, recovery, and handoff.
The team keeps unfinished patterns alive.
Risk is healthy when traces remain visible and recovery is possible.
The value of a system includes its ability to absorb mistakes, repair damage, and continue learning.
Critical Adoption
Use machines to widen thought, compare alternatives, extract patterns, and reduce repetitive effort.
Treat generated output as material rather than authority.
Adopt slop as raw ore, not final truth.
Slow down where understanding begins.
Humane Determinism
Machines may behave like functions.
Humans do not need to be described as functions.
Use deterministic tools without reducing people to deterministic terms.
Respect variation, interpretation, judgement, and context as essential parts of collaborative work.
Sustainable Attention
Protect attention from infinite formatting, planning, and production.
Prefer contact with reality.
Prefer demos over speculation.
Prefer testing over assumption.
Prefer workshop practice over simulation.
Capability grows where attention meets reality.
Goal
Build organizations that learn continuously.
Build systems that increase understanding.
Increase understanding to expand collective capability.
Use that capability to solve harder problems together.