Shared Artifact¶
Artifact — instantiates Productive Transition-Zone Design
A single object both regimes jointly own and read, plastic enough to mean something on each side yet fixed enough to coordinate them without meetings or translators.
Shared Artifact is the transition zone reduced to a thing: one durable object that sits in the overlap and belongs to both regimes at once. Its defining property is a productive double life — it is plastic enough that each side reads its own meaning into it, yet robust enough that both are held to the same commitments. That is what lets two regimes coordinate asynchronously, without a person in the middle or a meeting on the calendar: instead of understanding each other, each side only has to understand the artifact. Where a spanner network carries coordination in people and a studio produces it in events, the shared artifact stores it in an object that persists after everyone has gone home, becoming the zone's memory and its single source of truth for what crosses the boundary and in what form.
Example¶
A spacecraft's bus (power, structure, avionics) and its science instrument are built by two different teams, often two different organizations, on two different schedules. They cannot merge, and endless meetings won't scale. What holds them together is a single shared object: an Interface Control Document — the jointly owned specification of every place the two touch, down to the connector pinouts, the mounting-bolt pattern, the data rates, the thermal limits, and who is responsible for each.
Neither team needs to understand the other's internals. The instrument team reads, from the ICD, exactly how much power and what data pipe they will get; the bus team reads exactly what mass and heat they must accommodate. When the instrument's appetite grows, the change is negotiated on the artifact — a versioned edit both sign — rather than in a corridor conversation that one side later disputes. The ICD is the commons: the moment it goes stale and the real hardware drifts from what it says, the interface silently breaks, which is why its version history is guarded as carefully as the hardware.
How it works¶
- One object, two readings. The artifact is legible to both regimes but does not force either to adopt the other's worldview — each extracts its own obligations from the same shared thing.
- It fixes the interface, not the interiors. The artifact specifies precisely what crosses the boundary and in what form, and deliberately stays silent about how each regime meets that on the inside.
- Change happens on the artifact. Coordination is an edit to a jointly owned, versioned object, so agreements persist and drift is visible rather than living in fallible memory.
Tuning parameters¶
- Plasticity vs. rigidity — how much interpretive room the artifact leaves. Flexible artifacts adapt to each side's needs but let the two drift apart; rigid ones prevent drift but freeze the zone and invite gaming.
- Scope of what it fixes — a narrow interface contract or a sprawling shared model. Wider scope coordinates more but is costlier to keep true and easier to let rot.
- Ownership model — freely co-editable versus formally change-controlled. Open editing keeps it live; change control keeps it trustworthy but slow.
- Update cadence and versioning — how aggressively the artifact is kept in sync with reality, and how visibly. This dial decides whether the artifact leads the work or lags into fiction.
- Legibility to both sides — how much the format favors one regime's conventions. A format only one side reads fluently quietly hands that side control of the interface.
When it helps, and when it misleads¶
Its strength is coordination without co-presence: a good shared artifact lets two regimes stay aligned across time, distance, and turnover, and it offloads work that would otherwise fall on scarce human translators. It is also durable — the one part of the zone that remembers what was agreed after the people who agreed it have moved on.[1]
It misleads when the artifact and reality diverge and nobody notices — a stale spec is worse than none, because both sides trust it. It can be captured: whoever controls the format encodes their own worldview and makes the other regime a guest in it. Over-specification freezes a zone that should keep adapting, and "the document is the work" becomes theater where updating the artifact substitutes for doing the thing. The classic misuse is retrofitting the artifact after a decision to make it look jointly agreed. The discipline that guards against this is genuine joint ownership, strict version hygiene, and treating any drift between artifact and reality as a defect to be fixed on sight.
How it implements the components¶
Shared Artifact fills the shared-object components — the ones that live in a durable thing rather than in people or events:
shared_boundary_commons— the artifact is the commons: a single resource at the boundary that both regimes hold, maintain, and draw on in common.zone_specific_function_definition— by fixing exactly what crosses and in what form, the artifact inscribes the zone's specific coordinating function, giving the overlap a defined job rather than an open-ended mixing.
It does not supply the people who maintain and interpret it (the Boundary-Spanner Network), run the live recombination sessions its content feeds (the Cross-Functional Design Studio), govern rights and burdens over the commons (Joint Stewardship Council), or enforce what physically passes the boundary (Semipermeable Membrane).
Related¶
- Instantiates: Productive Transition-Zone Design — the artifact is the zone's coordination stored in a durable object.
- Sibling mechanisms: Cross-Functional Design Studio · Boundary-Spanner Network · Joint Stewardship Council · Semipermeable Membrane · Network Mixing Protocol
Notes¶
An artifact has no immune system of its own: left unattended it rots into a confident fiction. It therefore depends on a steward — typically the Joint Stewardship Council to govern it and the Boundary-Spanner Network to keep it honest against reality. Treat the artifact as an input to coordination, never as evidence that coordination has actually happened.
References¶
[1] In science-and-technology studies this is a boundary object (Star and Griesemer) — an object plastic enough to adapt to local needs on each side yet robust enough to hold a common identity across them. The dual "plastic yet robust" property named above is exactly what makes such an object work at an interface. ↩