π‘ The intention: from thought to execution
on November 15, 2026
An intention can be expressed in a few words: to understand, to traverse, to tell, to raise energy, to shorten the waiting time. It indicates what we are looking for. It does not yet say what a system can do, in what order, with which information or how we will know that the result is suitable.
It is this distance between the objective and the effect that the workshop From Intention to Execution proposes to explore. Its reading grid follows seven moments: intention, context, decision, planning or orchestration, execution, result and evaluation. It is not a universal theory. It is a way of making visible the choices that the word βautomationβ sometimes masks.
Darkwood serves here as a common ground: an ecosystem of services where an intention can become a game, a clip, a musical progression, a software operation or a business definition. The domains do not all share the same constraints, but they allow observing the same question from multiple angles.
An intention is not an action
Saying βI want to traverseβ does not move a character. One must know the ground, choose a direction, check that the rule allows this movement, and then accept its consequences. In software, the reasoning is comparable: the goal does not contain the available data, nor the constraints, nor the execution strategy.
The context gives the intention an operational meaning. A known route does not lead to the same decision as an unknown route. A scene that must preserve a character does not offer the same freedom as a scene of rupture. An input-output operation does not call for the same strategy as a heavy computation.
Darkwaar: deciding in a world of rules
Aftertrace provides a concrete example. The player may want to recognize a shore or to carry a ember. The world is the same, but the intention changes the value of each action.
When the map is unknown, exploring reveals safe squares, a crack, or a signal. Recklessly risking the ember does not necessarily answer the objective: the attempt may be technically possible and strategically wrong. When the route is known, the knowledge gained transforms the decision. The player can arrange their steps according to a route already understood.
The rule system thus separates several elements: the playerβs objective, the available information, the allowed moves, the steps consumed, and the final state. The evaluation asks not only βdid the movement work?β, but βdid the ember reach the exit under the expected conditions?β
The workshop takes up this distinction without modifying Aftertrace. Its local rules explain a choice; they do not reproduce the game engine.
The game Intention extends this idea with its dedicated mechanic: https://darkwoodcom.itch.io/intention
Heroclip: a creative intention constrained without determining everything
A narrative intention can aim at mystery, reunion or escape. Each influences the composition of the storyboard. Mystery delays the answer. Reunions bring trajectories together. Escape shortens idle times and accelerates decisions.
The creative context adds the constraints of continuity. Strict continuity preserves character, palette and prop. Visual continuity may preserve only the palette. Free direction allows a stronger rupture, at the cost of additional connections to verify.
In the studied Heroclip product, projects, generations, storyboards, lineage and bounded-context selection already provide concrete reference points. The workshop experience, however, remains fully deterministic: three cards are assembled locally, without calling a video model.
The technical result β three produced shots β does not guarantee narrative success. The revision must still ask whether rhythm, connections and the initial promise are perceptible.
The first public Heroclip clip devoted to intention makes this chain visible on a reconstructed artifact: https://heroclip.darkwood.com/clips/6e42bdb6-7ba2-42cf-91a1-cdb3b40b4fdc
DJ Matyo91: directing a musical progression
An artistic intention does not mechanically fix a sequence of sounds. It guides choices. An opening establishes reference points. A rise densifies the motif and progressively reduces the space between transitions. A peak seeks immediate impact while leaving room to avoid saturation.
The context here includes the place in the set, the available duration, the tempo and the sought energy. The workshop transforms these choices into an eight-step curve, a tempo and a transition proposal. A small Web Audio synthesis makes an original motif audible without samples or third-party tracks.
This demonstration is not tied to any unverified real set. It rather prepares an applicative field: make a musical intention explicit, compare several progressions and document why a transition is chosen.
The first public set by DJ Matyo91 uses only original sources and allows experimenting with playback, mixing and replay: https://djmatyo91.darkwood.com/sets/01a1dcb9-1d11-7b37-b596-57b95830bb3f
The Studio allows preparing these same actions in the replayable format: https://djmatyo91.darkwood.com/studio
Flow: composing and executing processing steps
Flow illustrates how a technical intention becomes a composition of treatments, then an execution strategy adapted to the context.
If the priority intention is to minimize operational complexity, inline PHP remains the natural choice: a runtime, a direct path and little supervision. If the objective is to reduce the waiting time of several independent input-output operations, Amp and Revolt allow overlapping them on a portable event loop. This strategy, however, requires suspendable operations and does not accelerate CPU calculations.
For CPU-intensive work, the honest decision is to measure isolation in workers or native kernels. The cost of transfer, the supervision and the actual gain are part of the result. The workshop provides no benchmark and presents no new backend validated in production.
Uniflow: From business intention to process
A business intention such as preparing a registration or an order becomes a more precise definition: collect the information, apply the rules, then prepare a confirmation.
The context can interrupt this sequence. If mandatory information is missing, the right decision is not to artificially produce a success, but to suspend before the effect and ask for the missing information.
Uniflow actually persists program definitions, their description, their data and their visibility. The local demonstration adds neither a durable engine, nor payment, nor a distant operation. It simply distinguishes the need, the rules and the produced definition.
Darkwood: From intention to the ecosystem
Darkwood links several services that transform an intention into creation, into an operation or into an observable experience. Darkwaar confronts the playerβs intention with the rules of a world. Heroclip translates it into narrative and visual choices. DJ Matyo91 guides it toward a musical progression and transitions. Flow composes and executes processing steps; Uniflow transforms a business intention into a process definition.
These services do not alone guarantee the success of the intention: they enable actions, preserve context and produce results that still need to be evaluated. The Memory Workshop completes this reading: memory preserves the information and states that illuminate the following decisions, while the intention gives them direction.
Technically succeeding may not be enough
A request can terminate without error and return the wrong information. A storyboard can be complete but miss its intention. A transition can be clean but break the desired progression. A movement can be approved while moving the player away from their objective.
Thus the evaluation compares the observable result to the initial intention. It makes the deviation explicit: result conforming, partial, contradictory or impossible in the current context. This deviation prepares a new decision rather than a blind repetition.
From Intention to Result
The intention provides a direction. The context gives it meaning. The decision selects a strategy. Planning organizes the steps. Execution produces effects. The result makes them observable. The evaluation finally says whether these effects meet what we were seeking.
Interactive Workshop: https://intent.content.darkwood.com/
Presentation: https://slides.darkwood.com/slides/intent