Six named arrangements of a team, and the level above it: one person working several teams at once.
Rev. 1 — 13 Aug 2026. First publication. Answers the concession made in OM-001's preface — that its acceptance checklist was the weaker thing Alexander warned about — and takes up the contention OM-002 §14 named and left open. Carries eight open evidence slots, marked as such in §4.
Rev. 2 — 15 Aug 2026. Retitled from Arrangements, and the Yard. The title is now its two sections exactly — Arrangements, and Many Teams — following the same rule as the rename of the catalogue itself. The URL, the number and every anchor are unchanged: /om-003 is what a citation hangs on. Ids published under the old pattern names still resolve.
Rev. 3 — 15 Aug 2026. Yard and fleet are both gone. Rev. 1 proposed yard while noting it was chosen rather than found and could be overturned cheaply by a later document; this is that overturning, and §14's aside now records it. The level is described plainly — several squads, worked at once — and no replacement term is coined. Squad stays, and OM-002 now uses it too. The section ids #the-yard and #can-an-agent-run-a-yard are unchanged: headings move, published anchors do not.
Rev. 5 — 15 Aug 2026. The argument moves out. §1 Two gaps, one of them our own and §5 Forces that recur are now OM-005 · On the First Four Papers §§11 and 13. §4, The evidence, which is weaker than OM-002’s, was split: the assessment went to OM-005 §12, and the four-source table with its eight open slots stayed here as §3, retitled The four sources, and eight open slots — twelve links from ten sections of this document point into those slots, eight of them from entries, and an entry should not have to leave its own paper to say what it rests on. #the-evidence therefore still resolves here, and all eight #slot-N anchors are untouched. The two retired ids are kept in §18. This document is now what it is called: six arrangements and three patterns, presented. Every section number in the body has been renumbered to match; numbers inside the dated notes above describe the structure of their own day and have been left.
Rev. 6 — 22 Aug 2026. Vocabulary and prose, under site/STYLE.md. squad becomes team throughout, 102 times, which finally makes the title, the section name and the prose agree: this paper was called Many Teams while its body said squad on every page. principal becomes person, 17 times. Two words used from the first page and defined nowhere are now defined in §1: a unit is a piece of work one agent can run to completion on its own, and a team is several agents working on one task for one person. All 97 uses of unit here are the house sense, checked one at a time rather than assumed. Em dashes fall from 133 to 32, and from 10.7 per thousand words to 2.6. The habit underneath them was worse. The phrase rather than appeared 63 times against 3 of instead of, a ratio near 21 to 1 where the Gang of Four's is close to even, and it is now 36 to 14. Most of that came from writing the plain form, not from swapping one phrase for another. Ten asserted virtues are gone; the ones left mark a boundary, a count, or a real state against an apparent one. §17's Named honestly, on the same principle as the two documents before it is gone with them, since a document that asserts its own honesty has spent something it should have been demonstrating. No id, name, relation or claim has changed, and no arrangement is described differently. The revision notes below, the sample transcripts, the quoted material and the eight #slot-N anchors are untouched.
Rev. 4 — 15 Aug 2026. One citation of OM-001 repointed: its acceptance checklist was §10 and is now §16, after OM-001 gave its fourteen entries their own numbered sections and moved its argument to OM-005. The Rev. 1 note above still cites /om-001#preface, which still resolves — that anchor is kept in OM-001 §18 and lands on the line naming where the preface went. Dated notes are records and are not rewritten. Nothing else in this document changed.
Two words carry over from OM-002. A unit is a piece of work one agent can run to completion on its own. A team is several agents working on one task for one person, arranged in one of the six shapes below.
The reason to name an arrangement is not tidiness. If a composite were only its constituents run in a particular order, the name would be a convenience and the honest thing would be to describe the order and move on.
A composite is worth a name when it changes what its constituent patterns mean. That is the test applied throughout this document, and it is not a low bar: one candidate failed it and was cut (§4).
Four examples, before any of the entries:
| Pattern | In plain fan-out | Inside another arrangement |
|---|---|---|
| Gather (OM-002 · 8) | Surfacing disagreement is a safety property. Ideally there is none, and finding some means something went wrong. | In a Panel it is adjudication, and disagreement is the entire product. A panel that agrees unanimously has told you almost nothing. |
| Isolation by Construction (OM-002 · 3) | Makes collision unrepresentable. It is about files, trees and devices. | Under Adversarial Verify it is about independence of judgement. The thing being kept apart is not two writes; it is two opinions, one of which would otherwise inherit the other. |
| The Return Path (OM-002 · 6) | Carries a report back to the context that commissioned the unit. | In a Pipeline the report is the next unit's commission. Two patterns collapse into one object, and the seam between them becomes the thing to get right. |
| Fire and Forget (OM-002 · 4) | Units proceed on stated assumptions, and the assumptions meet at the join where they may conflict. | In a Pipeline nothing meets. An assumption made at the first stage is inherited by the fourth, unexamined, because there is no second opinion anywhere in the line. |
Why this is the answer to a criticism, and not a taxonomy
Alexander told a room of object-oriented programmers in 1996 that the format had travelled to software and the important part had not. A pattern language was meant to be generative: a sequence you follow that produces a coherent whole, each step differentiating what the previous step laid down. What arrived instead was a shelf of separable good ideas, each individually defensible, with nothing saying how to go from one to a building.
OM-001 accepted the criticism and then demonstrated it. Its §8 turned fourteen patterns into fourteen questions to ask about a finished interface. That is useful, and it is the weaker thing: a checklist is applied afterwards, to something already made.
A catalogue of composites is the stronger thing, or the nearest this subject has come to it, for three reasons:
This is a partial discharge and should not be sold as more. Six named composites are still a shelf if nothing orders them, which is why §11 exists; and Alexander's real test was never structural. It was whether what gets generated is good for the people who have to live inside it. The person lives at the level above, which is the second half of this document, and the reason the two halves are one document.
Two levels, and they are not the same problem repeated. Below, a team takes one of several shapes, each an arrangement of OM-002's eight patterns. Above, a person works several such teams at once, and the constraints there come from the person rather than from the topology.
Level one · six arrangements
Level two · three patterns for many teams
Composites first, deliberately: you cannot describe what a person is orchestrating until there is a word for it, and what they orchestrate are arrangements. If the second half accumulates evidence of its own it becomes its own document, with a dated revision here saying so. The two halves are written to be separable for that reason.
Every entry below names which of four sources it rests on, and eight of them link to one of eight places where an incident belongs and none is on record. Both are here, with the entries that cite them.
OM-005 §12 compares this document’s evidence with OM-001’s and OM-002’s: thinner than either, and why that is said at the front rather than at the end.
What the sources are
| Source | What it can establish | What it cannot |
|---|---|---|
| Inherited measurements. OM-002's four console findings, unchanged and not re-derived here. | That the constituent patterns are real, and that their failure modes were observed, not imagined. | Anything about the arrangements. A measurement of a part is not evidence about the whole it sits in. |
| One practitioner's working practice, four to six areas at once, days rather than months, August 2026. | Existence and cost: that these arrangements get run, and that the failures named are cheap to hit. | Frequency, distribution, or whether other people's yards behave like this one. One practitioner is an existence proof and nothing more. |
| Reasoning about topology. What must be true of a shape given what OM-002 established about its parts. | Consistency, and predictions specific enough to be wrong. | Everything else. This is the largest single source in this document, and it is argument. |
| Adjacent prior art: pipelines, juries, ensembles, review, queueing (§19). | That the shapes exist and have been worked out elsewhere under real constraints. | That they transfer. The party being arranged here can decide to do something else, which none of those ancestors could. |
The open slots
Eight places where an incident belongs and none is on record. They are marked here, and again in the entry each one bears on, rather than being quietly absorbed into confident prose. A later revision fills them, dated, and says which claim moved.
Slot 1 · Pipeline
A run where a stage's intermediate artifact was not durable and an earlier stage had to be repeated: with what the repetition cost.
Slot 2 · Scout then Swarm
A plain fan-out in which units duplicated the same orientation work, measured: how much of each report was the same paragraph about the same layout.
Slot 3 · Loop Until Dry
A run stopped at a fixed round count, restarted, and found to have missed something: with what the extra round returned.
Slot 4 · Panel
One panel whose disagreement changed a decision, and one whose unanimity turned out to be worthless. Both are needed; the second is the harder to notice.
Slot 5 · Adversarial Verify
Confirmation rates with and without the producer's reasoning in front of the verifier. This is the one claim here that could be settled cheaply and has not been.
Slot 6 · Ordered by Readiness
How long a team sat blocked on a decision while another was being watched. The claim that this is the dominant loss at this level rests on it.
Slot 7 · Re-entry Note
The cost of re-entering a team cold, timed, against the cost of reading a maintained note.
Slot 8 · Landing Order
An instance of finished work invalidated by another team's landing: what was lost, and when it was noticed.
What follows from that
Every entry below carries an Evidence line saying which of the four sources it rests on. Where an entry rests on reasoning alone, it says so and points at its slot. This is not modesty as a rhetorical move: an entry that survives on topology alone is exactly the entry that should be argued out first if somebody's practice contradicts it.
Two kinds of entry, in one template. Composites (§5–§10) are arrangements of OM-002's eight patterns. Patterns (§13–§15) are new entries at this level, in the same form as OM-001's and OM-002's.
Citations: a composite is cited as Pipeline (Composite 2); a level-two pattern as Landing Order (OM-003 · 3); OM-002's eight as Gather (OM-002 · 8); OM-001's fourteen link to the canonical entry, as Honest Gauge (OM-001 · 12).
Two fields added, and why
OM-001 added one field to the Gang of Four template, Failure Signature, and named Alexander as the reason. This document adds two, for reasons it should have to defend:
The candidate that was cut
Long Runner with Check-Ins (one unit on a long horizon, reporting periodically) was on the list and is not an entry. It fails the test in §1. Its parts are Status Board (OM-002 · 5) and Fire and Forget (OM-002 · 4), and neither changes meaning inside it: the status board reports what the channel carries, exactly as before, and the check-in is the same tension between an interrupt and a report. Its failure signature is theirs, unmodified. It is also not an arrangement of units, since there is only one: a long run is a duration, not a topology.
The cut is recorded, not hidden, because the test is only worth anything if something fails it. The entry comes back, in a dated revision and with the argument attached, if a check-in turns out to change what Status Board means. That argument would have to be that a report the person will read hours later is a different kind of object from a surface they glance at.
Level one · Composite 1
Intent
Split work into units that never meet, run them at once, and reconcile their reports: the default arrangement, named so that the others read as departures from something rather than as inventions.
Motivation
This is the arrangement OM-002 describes end to end, and it is the right one often enough that it gets chosen by default rather than on purpose. Naming it costs nothing and buys the ability to say “this is not a fan-out” about the other five.
Its own failure is specific and common: work that does not partition, dispatched as though it does. Units then discover they need each other's results, proceed on guesses about them, and hand the whole problem to the gather, which is where the cost lands, all at once, in the one step that cannot be parallelised.
Applicability
When the units are genuinely independent: different files, different questions, different subsystems. When the number of units is known before starting. When each unit's result is useful on its own, so a failed unit costs one unit instead of the run.
Structure
What it changes
| Constituent | What it means here |
|---|---|
| All eight of OM-002 | Exactly what OM-002 says. This is the arrangement that document was written against, and the entry exists to make that visible, not assumed. |
An empty table would have disqualified any other candidate. This one is admitted as the baseline against which the tables below are read, and it is the only entry in this document with that exemption.
Participants
Commission is split before anything runs. Units are peers and do not communicate. Gather holds the decomposition and so knows where the units were expected to touch.
Collaborations
All eight of OM-002, in the order that document sets out: Proposed Dispatch (1), Authority at Dispatch (2) and Isolation by Construction (3) at the split; Fire and Forget (4) and Status Board (5) in flight; The Return Path (6), Nested Attribution (7) and Gather (8) at the join.
Consequences
Implementation
Decide the partition before dispatching and write down where units are expected to meet: that list is the gather's first search. Prefer units whose results are useful alone. When two units keep needing each other's output, that is not a coordination problem to solve with more messages; it is the wrong arrangement, and Composite 2 is the right one.
Failure signature
The same file in three units' reports. Or units that each proceeded on a guess about what another one did, producing a clean textual merge and a system that does not work: Gather's signature, arriving because the arrangement was wrong rather than because the gather was.
Evidence
Inherited measurements and practice. This is OM-002's arrangement, and its evidence is OM-002's.
Known uses
Map-reduce's map phase. make -j. A CI matrix. Any batch of independent jobs against a pool of runners.
Related patterns
Pipeline (Composite 2), Scout then Swarm (Composite 3), Gather (OM-002 · 8).
Level one · Composite 2
Intent
Run units in series, each one's report standing as the next one's commission, when the work does not partition but does decompose into stages.
Motivation
Some work has parts and no partition. Design, then implement, then test, then document. Survey, then rewrite. Each stage needs what the one before it produced, so dispatching them at once produces three units guessing at each other's output and a gather that has to invent the missing order after the fact.
The arrangement is old and the interesting part is what happens at the seam. The Return Path and the next Authority at Dispatch are the same object. That is not a convenience; it changes what a report is for. In a fan-out a report is an account for a human or a gatherer to read. Here it is an input to a machine that will act on it without discussion, which means an under-specified report does not produce a question; it produces a wrong stage.
The second change is quieter and is the reason this arrangement fails in ways fan-out does not. In a fan-out, units make assumptions independently, and the join is where incompatible ones collide and get caught. In a pipeline nothing collides. An assumption made in stage one is inherited by stage four unexamined, because there is no second opinion anywhere in the line. Errors do not conflict here; they compound.
Applicability
When each stage genuinely needs the previous stage's output. When the intermediate artifacts are durable and inspectable. Not applicable when stages are only conventionally ordered: “design then build” is a real dependency; “front end then back end” often is not, and pipelining it buys nothing but latency.
Structure
What it changes
| Constituent | In plain fan-out | Here |
|---|---|---|
| The Return Path (6) and Authority at Dispatch (2) | Two objects at opposite ends of the run. | One object. The report's shape is now a contract, because a machine acts on it directly. |
| Fire and Forget (4) | Assumptions are made in parallel and collide at the join. | Assumptions are made in series and compound. Recording them stops being hygiene and becomes the only audit there is. |
| Gather (8) | Reconciles N reports. | Nearly disappears: there is one lineage and nothing to reconcile. What remains is checking the last stage against the first commission, which is a different job. |
| Status Board (5) | How many are running, and what each is doing. | Which stage, and how far the line has moved. One unit runs at a time, so the useful number is position, not count. |
| Isolation by Construction (3) | Keeps concurrent writes apart. | Keeps stages recoverable. The value is that stage 2 can be re-run against stage 1's artifact without repeating stage 1. |
Participants
Stages run one at a time, in a fixed order. Seam is a report that is also a commission. Artifact at each seam is durable, so any stage can be re-run without repeating its predecessors.
Collaborations
Depends absolutely on The Return Path (OM-002 · 6), which stops being a convenience and becomes the mechanism. Leans on Assume or Ask (OM-001 · 5) at every stage, since a stage that stops to ask stalls the line instead of one unit. Nested Attribution (OM-002 · 7) is what lets a defect at stage four be traced to the stage that introduced it.
Consequences
Implementation
Fix the report's shape before the first dispatch and make every stage emit it: the seam is a contract, and a contract nobody wrote is discovered by breaking it. Keep every intermediate artifact durable and addressable, so re-running stage n does not mean re-running stage one. Carry assumptions forward explicitly in the report instead of in the work, since they are the one thing this arrangement will never catch on its own. Where a stage's assumption is load-bearing, the cheap remedy is to hang an Adversarial Verify (Composite 6) on that seam alone instead of abandoning the pipeline.
Sample interaction
seam · stage 1 → stage 2
artifact branch survey/console-naming (notes.md, 1 file)
assumed "shell" in IPC namespaces is load-bearing
refused did not touch the cargo feature name
commission stage 2 acts on notes.md only; may not widen scope
→ dispatched as unit 02 with this report as its commissionFailure signature
A person pasting one unit's output into the next unit's prompt: The Return Path's signature in series rather than in parallel. Or a defect at the last stage traced back three stages, each of which was correct given what it was handed. Or an early stage re-run because its output was not kept.
Evidence
Practice and adjacent prior art. The compounding-assumption claim is reasoning about topology and has no incident on record: slot 1.
Known uses
Unix pipes. Build DAGs. Compiler passes. Editorial workflows where a draft is commissioned, edited and then fact-checked by different hands.
Related patterns
Plain Fan-Out (Composite 1), Scout then Swarm (Composite 3), The Return Path (OM-002 · 6).
Level one · Composite 3
Intent
Spend one read-only unit on a survey whose product is the decomposition, then fan out over it: for work nobody present knows the shape of.
Motivation
Proposed Dispatch (OM-002 · 1) assumes a primary agent that already holds the context, and it is right whenever one exists. Often none does. The work is in a part of the system nobody has touched this month, and the person and the primary agent are equally ignorant of it.
Fanning out anyway has a specific and invisible cost: every unit spends its first minutes doing the same orientation. Five units read the same five files, form the same picture, and each writes a paragraph describing it. Nothing errors. The cost is paid five times and shows up as “agents are slow” rather than as duplicated work.
The remedy is a stage, not a pattern: one unit whose commission is to look and report, with no authority to change anything, followed by a fan-out whose partition is the scout's own map.
The map earns its keep twice. It is the decomposition, and it is also the expectation, so the gather at the end has something to check the results against, which a fan-out that partitioned by guesswork does not.
Applicability
When the shape of the work is unknown to everyone present. When the cost of a bad partition is high: overlapping units, or a gap nobody notices. Not applicable when the primary agent genuinely holds the ground: then the scout is a stage that costs latency and returns what was already known.
Structure
What it changes
| Constituent | In plain fan-out | Here |
|---|---|---|
| Proposed Dispatch (1) | The primary agent proposes, because it holds the context. | A dispatched unit proposes, because nobody holds the context. The person still approves: the pattern's pause survives, and its source moves. |
| Authority at Dispatch (2) | A scope wide enough to do the work, which is where the risk is. | For the scout, read-only: the cheapest scope there is, and one whose blind grant costs nothing. Scope stops being a compromise for one stage. |
| Gather (8) | Compares reports against each other and against a decomposition made by guess. | Compares them against a survey made by something that looked. It can now find a gap, not only a conflict. |
| Isolation by Construction (3) | Applies to every unit equally. | Asymmetric. The scout needs none (it writes nothing), which is why it can be dispatched before anything is decided. |
Participants
Scout reads and reports, and may not write. Map is a list of separable jobs, not prose. Swarm is a plain fan-out over that list. Gather holds the map as its expectation.
Collaborations
Composes Pipeline (Composite 2) and Plain Fan-Out (Composite 1), which is the first evidence in this document that composites nest. Realises Proposed Dispatch (OM-002 · 1) where no primary agent can. Its map is Constrained Choice (OM-001 · 6) made into a work list.
Consequences
Implementation
Commission the scout for a list, and say so in the commission: N jobs, each with the files it touches and what it must not touch. Give it read-only scope and nothing else, so it can be dispatched without a decision. Have the person approve the map instead of the units: that is the moment judgement is worth most, and it is one decision instead of N. Carry the map into the gather verbatim; a gather that only sees reports cannot notice a job nobody did.
Failure signature
Five reports containing the same paragraph about the same files. Or a fan-out where two units renamed the same thing differently because neither knew the other existed. Or a scout's output that a person has to turn into units by hand, which means the scout did half the job and handed back the half that does not scale.
Evidence
Practice and reasoning about topology. The duplicated-orientation cost has not been measured: slot 2.
Known uses
A survey before a refactor. git grep before splitting the work. Reconnaissance in the sense the word is borrowed from. A planning step whose deliverable is a list of issues rather than a document.
Related patterns
Plain Fan-Out (Composite 1), Pipeline (Composite 2), Proposed Dispatch (OM-002 · 1).
Level one · Composite 4
Intent
Repeat rounds of fan-out, deduplicating against everything seen so far, until some number of consecutive rounds return nothing new: for work whose size is not known before starting.
Motivation
Some work has no denominator. Find the bugs. Find every place this term is used incorrectly. Find the cases this parser mishandles. There is no list to partition because the list is what you are producing, and any fixed number of units is a guess about a quantity nobody has measured.
The guess is usually made once and never examined. Ten agents are dispatched because ten felt right; they return thirty findings; the run is declared complete. Whether there were thirty-one is unknown and unknowable from inside the run, and the tail (which is where the interesting findings live, because the obvious ones were found by everybody) is systematically missed.
Convergence has to be defined by the results instead. Run a round, deduplicate against everything seen in every previous round, and stop when consecutive rounds add nothing. The stopping condition is a property of the output, not a parameter.
The trap is precise and worth stating early because it is easy to build and hard to notice: deduplicate against everything seen, not against everything accepted. If findings that were examined and rejected are not in the seen-set, they will be rediscovered every round, every round will look productive, and the loop will never dry.
Applicability
Whenever the number of things to find is unknown: bug hunts, audits, sweeps for a pattern across a codebase, edge-case discovery. Not applicable where the work is enumerable: if the list can be computed, computing it is cheaper and complete.
Structure
What it changes
| Constituent | In plain fan-out | Here |
|---|---|---|
| Gather (8) | A final step that reconciles N reports once. | A step between every round, whose main job is deduplication and whose output feeds the next round's commissions. It runs K times and its state persists. |
| The Return Path (6) | Delivers a report to the commissioning context. | Must be durable across rounds. A report that dies with its round breaks convergence instead of losing information, because the seen-set is the stopping condition. |
| Isolation by Construction (3) | Isolates units from each other. | Also isolates rounds from each other, so that a round's finding is attributable to that round, which is how you know the loop is drying rather than repeating. |
| Proposed Dispatch (1) | The decomposition is approved once, before the run. | Each round is a new decomposition, and asking for approval every round destroys the arrangement. Approval moves to the loop's terms (the stopping rule and the budget), not its units. |
Participants
Round is a fan-out. Seen-set holds every finding ever surfaced, accepted or rejected. Stopping rule is stated before the first round. Budget is the other stopping rule, and the honest one.
Collaborations
Contains Plain Fan-Out (Composite 1) as its round, and usually Adversarial Verify (Composite 6) as its filter: findings must be judged before they enter the kept set, though they enter the seen-set either way. Honest Gauge (OM-001 · 12) governs how “dry” is reported: two empty rounds is a stopping rule, not a proof of completeness, and a surface that says “all findings” has lied.
Consequences
Implementation
State the stopping rule before the first round (K consecutive empty rounds), and state a budget alongside it, because the rule alone does not terminate. Key the seen-set on something stable enough to match near-duplicates: the same finding described in two ways is the common case, and an exact-string key will never dedupe it. Tell each round what is already known, or it will spend its budget rediscovering it. And report the rule with the result: “two consecutive rounds returned nothing new” is a true sentence, and “we found everything” is not one this arrangement can support.
Sample interaction
sweep · incorrect uses of "shell" in prose
rule stop after 2 consecutive rounds with nothing new
budget 8 rounds
round 1 5 units +7 new (7 seen)
round 2 5 units +2 new (9 seen, 3 rejected — still seen)
round 3 5 units +0 dry 1/2
round 4 5 units +0 dry 2/2 → stop, 6 keptFailure signature
A team size chosen because it felt right, and a report that says “we found N” with no account of what would have found N+1. Or the same three findings surfacing in every round, with the loop apparently productive and never converging. Or a run declared complete after one quiet round.
Evidence
Practice and adjacent prior art. The claim that a fixed count systematically misses the tail is reasoning, and the run that would demonstrate it is slot 3.
Known uses
Fuzzing to coverage saturation. Capture–recapture sampling in ecology, whose entire purpose is estimating what has not been seen. Bug bashes run until the finding rate falls. Search that stops when a page of results adds nothing.
Related patterns
Plain Fan-Out (Composite 1), Adversarial Verify (Composite 6), Gather (OM-002 · 8).
Level one · Composite 5
Intent
Give several units the same question, kept apart on purpose, and make the disagreement between their answers the product rather than a problem.
Motivation
Every arrangement so far divides work. A panel duplicates it (the same question to each unit, no partition at all), and that inversion changes what the join is for.
In a fan-out, disagreement between reports is a warning: two units have contradicted each other and something must be reconciled before it reaches the artifact. In a panel, agreement is the uninformative outcome and disagreement is the finding. Where three independent answers diverge, you have learned that the question is underdetermined, which is worth more than any one of the answers.
That only holds if the units are genuinely independent. Let them see each other's drafts and they converge: not because the question was settled but because the channel carried an answer. Unanimity produced by a shared channel is indistinguishable from unanimity produced by a clear answer, and only one of them is worth anything. This is why isolation, which in a fan-out is about files, is here about judgement.
The other honest limit: a panel is for questions no test can settle. Convening three units to decide whether a function returns the right value is money spent to reproduce something a test would have told you for free.
Applicability
Questions of design, approach, naming, risk, or interpretation, where there is no oracle and being wrong is expensive. Where the units can be kept independent. Not applicable to anything checkable, and not applicable where the disagreement will be averaged, since averaging discards exactly what the arrangement was bought for.
Structure
What it changes
| Constituent | In plain fan-out | Here |
|---|---|---|
| Gather (8) | Reconciliation. Disagreement is a hazard to be surfaced and resolved. | Adjudication. Disagreement is the product, and a panel that reports none should have to say how hard it looked. |
| Isolation by Construction (3) | Keeps writes apart. Its subject is a tree. | Keeps judgements apart. Its subject is a channel, and it is violated by a helpful summary rather than by a concurrent write. |
| Authority at Dispatch (2) | Bounds what a unit may change; the risk is the action. | Usually read-only. The risk is the answer, so the scope is cheap and the attention moves to the question's wording, which every unit shares. |
| Nested Attribution (7) | Lets a result be traced to an actor. | Load-bearing, not hygienic: an opinion whose author is unknown cannot be weighed against another, and a merged panel is one voice pretending to be three. |
Participants
Question is identical for every unit, and its wording is now a shared dependency. Members are isolated from each other for the duration. Adjudicator applies a rule fixed before the answers arrived. Disagreements are an output in their own right.
Collaborations
Inverts Plain Fan-Out (Composite 1): same work to every unit rather than different work to each. Distinct from Adversarial Verify (Composite 6), which is asymmetric: a panel answers a question, a verifier attacks an answer. Depends on Nested Attribution (OM-002 · 7) to keep the members' voices separate.
Consequences
Implementation
Give every member the same question, verbatim, and vary nothing but the seat: a panel whose members were asked slightly different questions cannot disagree meaningfully. Keep them from each other for the duration, including from each other's artifacts; the leak is usually a shared workspace or a helpful summary, not a message. Fix the adjudication rule before the answers arrive, because a rule chosen afterwards is a preference with a quorum. Report the disagreement in the output, not only the verdict. And if a test could answer the question, run the test.
Sample interaction
panel · should the crate keep the name "shell"?
rule 3 members, independent; verdict needs 2; dissent always reported
member 1 rename — the word collides with the OS sense
member 2 rename — but keep the IPC namespace, it is load-bearing
member 3 keep — the collision is theoretical, the churn is not
verdict rename (2 of 3)
dissent member 3 names a cost the others did not priceFailure signature
Unanimity you cannot account for. Or a panel report that reads as one voice, so no reader can tell whether three units agreed or one summarised. Or a decision made by panel that a five-line test would have settled.
Evidence
Practice and adjacent prior art: juries and review panels have worked out the independence rules over centuries. The claims here about what leaked context does are reasoning; slot 4 is what would settle them.
Known uses
Juries, and their rules against discussing the case outside the room. Blind peer review. Ensemble methods, where the diversity of the members is the source of the gain. Design reviews with independent written positions collected before discussion.
Related patterns
Adversarial Verify (Composite 6), Plain Fan-Out (Composite 1), Isolation by Construction (OM-002 · 3).
Level one · Composite 6
Intent
Commission units to refute a result rather than to review it, with the burden of proof on the claim and a rule fixed before the verdicts arrive.
Motivation
A unit finishes and reports success. The report is coherent, the reasoning is legible, and it is wrong. Nothing in the fan-out will catch this: the gather compares reports against each other, and a single confident report has nothing to contradict it.
The obvious remedy, asking another unit to check it, fails in a specific way. Handed the claim and the reasoning, a checker reads the reasoning, finds it plausible, and agrees. It has been given the conclusion and a path to it, and confirming is the cheapest thing it can do. What comes back is a confirmation whose information content is close to zero, and which now reads as independent corroboration.
Two changes fix it, and both are asymmetries. The verifier's commission is to refute, not to check: the default outcome is that the claim fails, and surviving is the exception. And what it is given is the claim and the evidence, not the producer's reasoning, because reasoning is contagious, and a verifier that inherits it is not a second opinion but an echo.
This is where Isolation by Construction (OM-002 · 3) stops being about files. What must be kept apart is not two writers but two judgements, and the boundary is informational, not physical.
Applicability
When a claim is expensive to be wrong about and cannot be checked mechanically. When a run's findings will be acted on without further review. Not applicable when a test exists, and not applicable to everything indiscriminately: verification costs as much as production, and verifying every result doubles the run.
Structure
What it changes
| Constituent | In plain fan-out | Here |
|---|---|---|
| Isolation by Construction (3) | Prevents collisions between writes. | Prevents one judgement inheriting another. The boundary is informational, and it is crossed by sharing a rationale, not by sharing a directory. |
| Authority at Dispatch (2) | Says what a unit may do. | Also says what a unit may know. This is the first place a commission's contents matter as much as its permissions. |
| Gather (8) | Reconciles reports into an account. | Applies a decision rule to verdicts. It is counting, and the counting rule must predate the count. |
| Fire and Forget (4) | Units proceed instead of asking the person. | Refuters must not ask the producer either. The question they would ask is exactly the reasoning being withheld. |
Participants
Producer emits a claim and the evidence for it. Refuters receive both, and not the rationale. Rule is fixed in advance and stated. Verdict defaults to refuted where a refuter is uncertain, because the burden sits on the claim.
Collaborations
Usually the filter stage inside Loop Until Dry (Composite 4), and a good repair for a load-bearing seam in a Pipeline (Composite 2). Distinct from Panel (Composite 5) by asymmetry: a panel has no incumbent answer, and here the answer is the thing under attack. Sharpens Progressive Autonomy (OM-001 · 11), since a producer whose claims survive refutation has earned something measurable.
Consequences
Implementation
Word the commission as refutation: “find the input on which this is wrong”, not “review this”. Send the claim and the evidence; withhold the rationale, and check what the commission carries rather than what it was meant to. Default an uncertain refuter to refuted, since the point is to make survival mean something. Fix the rule and the number of refuters before dispatch. And prefer several verifiers with different angles (correctness, cost, does it reproduce) to several identical ones, because redundancy catches only what one lens can see.
Failure signature
Everything verifies. A confirmation rate near one is not evidence that the work is good; it is evidence that the verification is not working, and it is the cheapest thing in this document to measure. Or verifiers whose reports restate the claim approvingly. Or a verdict rule that changed once the verdicts were in.
Evidence
Reasoning about topology and adjacent prior art. The confirmation-rate comparison that would settle it is slot 5, the one open slot here that could be closed in an afternoon.
Known uses
Red teams. Adversarial review in security. Falsification in experimental design, where a hypothesis earns standing by surviving attempts to kill it. A defence that never sees the prosecution's reasoning, only its evidence.
Related patterns
Panel (Composite 5), Loop Until Dry (Composite 4), Isolation by Construction (OM-002 · 3).
Six named arrangements are a shelf unless something says how to choose. The choice is a short sequence of questions, asked in order, each one eliminating arrangements rather than scoring them.
Two things the figure cannot say. Arrangements nest, and the sequence is applied at each level: Scout then Swarm is a pipeline whose second stage is a fan-out; Loop Until Dry is a loop whose body is a fan-out and whose filter is usually an Adversarial Verify; a Panel sits comfortably inside a pipeline stage where one decision has to be made well. Ask the questions again at each level and the answers differ.
And the arrangement is a decision the person makes, not the primary agent. Proposed Dispatch (OM-002 · 1) hands the decomposition to the party with the most context, and it is right about the units. The shape is different: it encodes how much the person is willing to spend, how expensive being wrong is, and how much they trust the answer: three things the primary agent does not know.
What this is worth against the criticism
Honestly scored: this is more than a checklist and less than what Alexander meant.
It is more, because it runs forwards. Five questions asked before the work produce an arrangement, and the arrangement says which patterns apply and what each one is for at that point, where OM-001 §16 asked fourteen questions about something already built.
It is less, in two ways worth stating rather than glossing. Alexander's sequences unfold a single whole, each step differentiating what the last one laid down; this is a classifier that selects a shape and then hands over. And his final test is not structural at all: it is whether the thing generated is good for the people who have to live inside it. Nothing at this level can answer that, because at this level nobody lives anywhere. The person is one level up, and that is where the rest of this document goes.
One person, four to six areas of a product, each with its own dispatcher and its own units. The obvious move is to read that as OM-002 applied one level up: the person is a dispatcher, the teams are units, and the eight patterns should recur. They do not, and the reason is a single asymmetry that nothing at this level can remove.
On the word. Rev. 1 of this document proposed yard for the level above a team, while saying in the same breath that it was chosen rather than found and that a later document might overturn it cheaply. Rev. 3 takes that option. The word is gone and the level is described plainly instead: several teams, worked at once, by one person. Nothing below rested on it, which is exactly what made the removal cheap; the earlier text said so, and it was right. OM-002's fleet went the same way and for the same reason: both were containers wearing concrete clothes, asking a reader to learn a private vocabulary before reaching a single pattern. The level below kept its own name through that pass and lost it in Rev. 4, for the same reason: team is the ordinary word, and OM-002 uses it too.
The person cannot be commissioned
In a team, the dispatcher is an agent, and that is what makes OM-002's patterns possible. You can bound what it may do. You can give it a return address that resolves. You can require a report and read it. Across teams, the dispatcher is a person, who has none of those properties: not incidentally, but by construction.
| Dispatcher in a team | Dispatcher across teams | |
|---|---|---|
| Authority | Scoped at dispatch and enforced by the environment. | Unscopable. A person may always do the thing; “may not” is advice, and the environment they act in is the one that grants everything. |
| Addressing | A return address that always resolves. Delivery is observable. | Best-effort. You cannot route to a person's attention on demand, and the acknowledgement is somebody noticing, which nothing reports. |
| Accountability | Held to a report, which can be compared with the commission. | Writes no report and answers to nothing in the system. The record of what was decided exists only if the person chose to write it. |
| Scarce resource | Compute, which is purchasable and divides. | Attention, which is neither. Adding teams does not add capacity to look at them. |
| Failure mode | A unit that did not report, which is visible as such. | A team nobody looked at, which is indistinguishable from a team that is running quietly. |
Everything difficult at this level falls out of that table. Attention scheduling: which team you look at now is the only real decision, and nothing schedules it. Re-entry cost: switching teams means reconstructing a state that lives in a head, and it is paid on every switch. Contention isolation cannot solve: five teams, one repo, one CI, one merge target — Isolation by Construction (OM-002 · 3) copies workspaces and cannot copy the trunk. Landing order: the order finished work lands is a decision no unit can make, because each knows only its own. Uneven readiness: teams finish when they finish, and the queue is never empty and never full.
Three of those have entries below. The others are named here and left as forces, because a pattern needs a stateable failure signature and these do not yet have distinct ones.
Note on this half's evidence
One practitioner, four to six areas, days rather than months, and no instrument that measures where attention went. Every claim below is either an existence proof or reasoning; none is a frequency. Slots 6, 7 and 8 are the incidents that would change that.
Level two · Pattern 1
Intent
Order the surface across teams by what needs a person, not by what is happening, so the scarce resource is spent on decisions rather than on scanning.
Motivation
Status Board (OM-002 · 5) answers what is happening, and at team scale that is the right question: the person wants to know whether to intervene. Across teams it is the wrong one. Four teams are always happening. The question is which of them needs the person next, and for what decision.
The distinction that makes this a separate pattern rather than the same surface scaled up: readiness is not status. A team that is done is not necessarily ready to land, and a team that is running may have been blocked on a thirty-second answer for six hours. A surface ordered by status buries the second behind three teams that are working fine, and the loss is invisible because nothing failed.
The states that matter at this level are about the person, not the work: blocked on you, ready to land, running, finished and waiting, rotting. Only the first demands attention now; the second demands a decision soon; the rest demand nothing, and a surface that presents all five with equal weight has made the person do the sorting.
Applicability
Whenever more than about two teams run at once. Especially where teams are long-lived and the person's attention is the constraint, which at this level it always is.
Structure
Participants
Teams report their own readiness, which they know and the surface does not. Readiness orders by the decision being asked for. Person works down the list and stops when nothing is asking.
Collaborations
Sits above Status Board (OM-002 · 5) and inverts its sort key. Depends on Honest Gauge (OM-001 · 12): “blocked on you” must be wrong only when the system is wrong, or the list stops being read. Ambient Activity Channel (OM-001 · 9) is how the one state that matters (a team newly blocked) reaches a person who is deliberately looking elsewhere.
Consequences
Implementation
Sort by the decision being asked for, and show the age of that decision: a team blocked six hours reads differently from one blocked six minutes, and the age is the number that changes behaviour. Distinguish done from ready to land; they are different states and conflating them is how finished work sits unlanded. Let the list be empty, and make an empty list mean something: nothing here needs you. Keep the team-level surface underneath instead of replacing it, since once a team is chosen the question reverts to what is happening.
Sample interaction
4 squads
presence blocked on you 6h12m "rename the capability or keep it?"
patterns ready to land 2 files, no conflicts with console
console running 3 units, last event 40s ago
voice running 1 unit, quiet 22m — no news, not no progressFailure signature
A team that sat blocked on a small decision while the person watched one that was fine. Or the person opening each team in turn to find out which one needs them: Status Board's signature one level up, with the difference that here a status surface is what produces it rather than what cures it.
Evidence
Practice, and reasoning from the asymmetry in §12. The waiting time that would make the loss concrete has not been measured: slot 6.
Known uses
A code review queue sorted by “waiting on you”. Air traffic control strips, which are ordered by what needs a decision next. Kanban boards where the blocked column is read first. Triage in any sense of the word.
Related patterns
Landing Order (OM-003 · 3), Status Board (OM-002 · 5), Honest Gauge (OM-001 · 12).
Level two · Pattern 2
Intent
Have each team maintain a short note written for the person's return, so that coming back to it costs reading rather than reconstruction.
Motivation
The cost of running several teams is not the running. It is the switch. Returning to a team after a day away, the person has to rebuild: what this area is for, what was decided and why, what it is waiting on, what will break if it lands after the other one. None of that is written down anywhere in that form. The transcript is a record of how the state was reached, which is the expensive way to learn what it is. The diff says what changed, not what was intended.
So the person reconstructs, and the reconstruction is the real cost of running several at once: paid on every switch, invisible because it looks like working, and rising with the number of teams.
This is Durable Memory (OM-001 · 14) with the memory belonging to the wrong party. There, an agent keeps what matters so the person need not restate it. Here the reader is the person, and what must survive is the state of an area they were absent from, which means the note has to be written by the team, for a reader who was not there, and maintained, not composed once.
The obvious objection is right and is the pattern's main cost: a second artifact can go stale, and a stale note is worse than no note, because it will be believed. That makes this an application of Honest Gauge (OM-001 · 12) as much as of memory: a note that is not maintained must be visibly not maintained.
Applicability
Whenever the gap between visits to a team is long enough that state must be rebuilt: in practice, more than a few hours. Not applicable to a team being worked continuously; there the note is overhead and the working context is the memory.
Structure
Participants
Gatherer writes the note, because it already holds the decomposition and has just read every report. Note has a fixed shape and is short enough to read in full. Person reads it on return and is not expected to read anything else first.
Collaborations
An application of Durable Memory (OM-001 · 14) where the reader is the person. Written at the end of Gather (OM-002 · 8), which is the only moment the whole state is in one place. Feeds Ordered by Readiness (OM-003 · 1), whose one-line summary of a team is the note's first field. Governed by Honest Gauge (OM-001 · 12), because a stale note is a green light over a broken system.
Consequences
Implementation
Fix the shape and keep it short: what this team is for, where it stands, what it is blocked on, what has been decided and when, and what must land before it. Write it at the end of every gather, as part of the gather instead of as a chore after it. Date every line that is a decision, since the ones without dates are the ones that will mislead. Keep it in the team's own workspace so it travels with the work. And prefer replacing lines to appending them: a note that only grows is a transcript with extra steps.
Sample interaction
re-entry · presence
for the ambient light channel and its states
state 2 units done, 1 blocked
blocked "capability" vs "affordance" in the public API — needs you
decided 13 Aug · states are named, not numbered (breaks the old client)
lands after console/docs-rename, which renames the same symbolsFailure signature
Reading a diff to remember why a team exists. Typing “where were we?” to an agent. The same context explained twice in a week to the same team. Or a decision re-made differently, because the first one was in a transcript nobody reopened.
Evidence
Practice. The comparison that would give the claim a number (reconstruction timed against reading) is slot 7.
Known uses
Handover notes at a shift change, in every trade that has shifts. A ship's log. The STATUS file at the top of a long-running branch. A good issue description, maintained, not written once.
Related patterns
Ordered by Readiness (OM-003 · 1), Durable Memory (OM-001 · 14), Gather (OM-002 · 8).
Level two · Pattern 3
Intent
Decide the order in which finished work lands, explicitly and before it is finished, because the shared trunk is the one thing isolation cannot copy.
Motivation
Isolation by Construction (OM-002 · 3) makes parallel work possible by giving every unit its own copy, and it says plainly that conflict is deferred rather than removed: deferred to the join, where a person can see it. At this level, every team's join arrives at the same place: one repository, one CI, one main branch, one release, one production database. Isolation copies workspaces. It cannot copy the trunk.
Three things follow, and none of them is a team-level problem.
Finished work rots while it waits. A branch that is correct on Tuesday is a rebase on Thursday, because the ground moved under it. The cost of landing rises with the time held, so holding finished work for a coherent release is not free; it is a loan against a rising rate.
A landing changes other teams' ground. Team B's finished work can be invalidated by team A landing first, and nobody is informed, because A's units know only A's work. The failure surfaces as a CI break in a branch that nobody touched.
And the order is a decision no unit can make. Each team knows its own readiness and nothing about the others; the ordering information exists only in the person, which is precisely the party with no capacity to spare. Uneven readiness guarantees the queue is never empty.
Applicability
Whenever two or more teams land into the same target. The pattern is worth its cost from the second team onwards, and by four it is the dominant constraint at this level.
Structure
Participants
Teams produce finished work at different times. Trunk is the one target, and the list of things like it (CI capacity, the release, the production database, a token with one permitted caller) is the set of singletons shared across teams. Order is stated at commission time and revised out loud. Person owns it, because nothing else can.
Collaborations
Receives what Isolation by Construction (OM-002 · 3) defers, at the level where the deferral finally has to be paid. Depends on Ordered by Readiness (OM-003 · 1) to distinguish done from ready to land. A change of order is a dispatch to the teams whose ground moved, not a note. That dispatch is a Authority at Dispatch (OM-002 · 2) like any other.
Consequences
Implementation
State the intended order when the teams are commissioned: it is part of the decomposition, not an afterthought, and it is cheapest to get right before anything is built. Keep a written list of the shared singletons: the trunk, CI capacity, the release channel, the production database, the domain, any token with one permitted caller, and the person. Those are what cannot be isolated, and everything that contends is contending for one of them. Prefer landing small and often, and treat a long-held branch as a cost that is accruing, not as work that is safe. When the order changes, dispatch that fact to the teams it moves, and expect them to re-verify, not assume.
Sample interaction
landing queue · 13 Aug
1 patterns/name-states ready no shared symbols
2 console/docs-rename ready renames what presence depends on
3 presence/capability-api blocked re-verify after 2 lands
— voice/streaming-fix running lands last, touches the trunk widely
note: 2 before 3 was decided at commission time, not at merge timeFailure signature
Two finished branches waiting on each other while a person decides from memory which goes first. Or a team's work invalidated by another's landing, discovered by CI on a branch nobody touched. Or a rebase that costs more than the work it is rebasing.
Evidence
Practice, and reasoning from the singleton list. The invalidation incident that would make the cost concrete is slot 8.
Known uses
Merge queues and landing trains. Release trains with a fixed cadence. Work-in-progress limits in kanban, which exist because finished-and-unlanded work is the expensive kind. A slipway, which admits one hull at a time regardless of how many are built.
Related patterns
Ordered by Readiness (OM-003 · 1), Isolation by Construction (OM-002 · 3), Gather (OM-002 · 8).
This is stated as open because it is open, and because the alternative (answering it from taste) is how a pattern language acquires entries that are preferences wearing the template.
The case for is straightforward. Everything in §13 to §15 is scheduling, bookkeeping and ordering: readiness, notes, a queue. Those are exactly the tasks that survive delegation well, and the person's own account of what they contribute is not any of them:
“The whole time, you are giving the thing that the agents need, which is the taste and the direction and the choices that make this thing you're building yours.”
Working notes, 13 August 2026 — quoted in OM-002 §3
The case against is the asymmetry. An agent running this level would need authority over the singletons — the trunk, the release, the order in which things land — which is the one place where Authority at Dispatch (OM-002 · 2) is hardest to write, because the dangerous actions are exactly the ones the job consists of. It would need a return path to a person that tolerates the person not answering, which is a channel nobody has built. And it would have to hold four to six areas well enough to know that landing A invalidates B — a judgement that currently rests on knowing why B exists.
The observation that makes this more than an open question: the asymmetry does not disappear when you delegate this level. It moves. Put an agent in charge and the person becomes the person of that agent, and every property in §12's table reappears one level up — unscopable authority, unreliable addressing, no report, attention as the constraint. The boundary is not a level in the hierarchy. It sits wherever the person is, and it moves when they do.
What would settle it, in the direction of yes: this level run by an agent for a week, whose landing order produced fewer invalidations than the person's, with the person still supplying direction. In the direction of no: the same trial producing more invalidations, or producing the same number while the person spent as long supervising as they had spent deciding. Neither has been run.
What can be said now is narrower and worth saying: an agent can hold this level's state without holding its authority. Readiness, the notes, the queue and the singleton list are all things a team or a level-two agent can maintain, and maintaining them is most of §13 to §15. Whether the ordering decision itself can move is the open part.
There is more to name here than in the two documents before it, because there is less evidence.
Rev. 1 of this document opened with the two gaps it fills, an assessment of its own evidence, and the forces its entries trade off. Rev. 5 moved them to OM-005 · On the First Four Papers so that this document could be what it is called: six arrangements and three patterns for many teams, presented.
Each id published here is kept and lands on its own line below. The rule is mechanical: /om-003#x became /om-005#om-003-x. One of the three was split rather than moved, and that is the exception worth reading.
#the-evidence did not move, and still resolves here. That section was doing two jobs under one heading. The assessment (that this document's evidence is thinner than either of its predecessors, and the ladder showing how much) is OM-005 §12. The instrument (the four sources with what each can and cannot establish, and the eight open slots) stayed as §3, because twelve links from ten sections of this document point into those slots, eight of them from entries, and an entry should not have to leave its own paper to say what it rests on. All eight #slot-N anchors are untouched.
Directly ancestral
Adjacent traditions
Primary material
One practitioner's working practice, August 2026: four to six areas of a product run at once, each with its own dispatcher, over days rather than months. The measurements quoted anywhere in this document are OM-002's and are not re-derived here. No new instrument measurements were made for this document, which is the single most important thing to know about it, and §3 lists the eight incidents that would change that.
What is claimed
Less than OM-002 claimed, which was already less than OM-001. Fourteen patterns with four measured; then eight with two measured; now six composites and three patterns with none newly measured, resting on practice, on prior art in fields whose parts could not decide to do something else, and on reasoning about topology.
It is published in this state on the same principle as its predecessors: an argument in public gets corrected faster than one held back until it is safe, and the open slots make the corrections cheap to aim. The composites are the more confident half. The second half is a first description of a level that will look different once anybody else describes theirs.
Corrections and counter-examples: hello@organon.art. An arrangement shown not to change what its parts mean should be cut, as one already has been.