How a Session Works: Version 0.1

Every creative system needs a heartbeat. A moment that says: this is where it begins. For Jerry Gretzinger, that moment is walking to his table and drawing a card. The card decides. He follows.

My version is a little more elaborate — phases, rules, numbers, a hierarchy of decisions — but the spirit is the same. Sit down. Roll the dice. Find out what today's work is going to be.

Let me walk you through it.

Sessions, Phases, and the Shape of a Working Day

The session is the fundamental unit of this project. Everything happens inside one. Each session is logged, tracked, and filed away so that the history of the map is always recoverable — every decision, every tile, every number that drove a choice.

In this version of the rules — and calling it version 0.1 feels about right — a session is divided into four phases. They run in order, each one feeding into the next.

Initiation sets the scene. The dice are rolled here, and those numbers become the engine that drives everything else in the session. Initiation also selects the tile to be worked on — or decides that a new tile needs to be born. Eventually this phase will also handle tile rotation and repositioning on the map, but those rules do not exist yet. The system grows as it needs to.

Preparation is where the physical groundwork happens. A new tile gets its base coat and its metadata written on the back — coordinates, session ID, date, orientation. The quadrant cross is drawn, the anchor dot placed. The tile goes from a blank hexagon to something with identity and position in the world.

Tile work — still searching for a better name for this phase, suggestions welcome in the comments — is where the actual making happens. The dice determine what substrate or base material goes onto the tile: what kind of paper is glued down, how it is treated, what marks or details get added. At this stage of the project, that mostly means paper and glue. But the same phase will later handle paint, collage, ink, whatever the map demands as it grows.

Non-map activities are governed by the special die and can mean almost anything: preparing cards for the deck, writing a blog post, scanning completed tiles, printing backups, updating the website. This post, for instance, exists partly because the dice said it was time to share something. The system earns its keep.

The Initiation Phase, Up Close

So. The session starts. The dice are rolled. Six numbers appear on the table.

The first thing to establish is where the last session left off — specifically, which tile was the previous target. That tile's coordinates become the anchor point for calculating where to go next.

The coordinate system uses two axes: NE-SW and SE-NW. Negative numbers sit to the left, positive to the right, and the very first tile placed — T001 — lives at [0, 0], the origin of everything. It looks unintuitive at first glance, but it becomes natural quickly. The key thing is that every tile has a precise, unambiguous address.

The new target coordinate is calculated directly from the dice. The NE-SW value is blue minus green. The SE-NW value is red minus yellow. This can produce a wide range of results — and it can absolutely send the target to the same tile twice in a row, which is fine. What it can also do, over many sessions, is pull the map in one direction while leaving other areas untouched for a long time. That is something worth watching. A solution may be needed eventually. For now, the map goes where the dice send it.

So: a new target coordinate exists. But there may not be a tile there yet.

Finding the Target: Adjusting the Aim

When the calculated coordinate is empty — no tile exists there — the system does not simply create one on the spot. Instead, it walks the target back toward the existing map, one step at a time.

The logic works like this. From the empty coordinate, identify the nearest of three reference lines running through the origin: the diagonal where the absolute values of both coordinates are equal, or either of the two axes where one coordinate is zero. If the target is equidistant between two of these lines, the black die breaks the tie — odd sends the target to the counter-clockwise option, even to the clockwise one.

From there, move one step toward that reference line. Check the new position: is there a tile there? Is it at least adjacent to an existing tile? If not, move again. Keep moving until the target lands somewhere that touches the existing map.

To make it concrete: a target at [-4, 5] would first identify the diagonal as its nearest reference line. It steps to [-4, 4], then along the diagonal toward [-3, 3], [-2, 2], [-1, 1]. At [-1, 1], the target is adjacent to an existing tile. Stop.

Now the brown die makes the final call. If it is even, a new tile is created at that location and the session continues from there. If it is odd, the target moves one step further — which might land on an existing tile, or might edge to another empty space that is adjacent to the map. Either way, the brown die is what decides whether something new is born or whether the session revisits somewhere that already exists.

It is a simple mechanism with surprisingly interesting consequences.

Logging the Session

The target is chosen. Whatever the map holds from this point forward is in the hands of the dice. Before moving into the preparation phase, the session log gets its first entry.

At this stage, the log records:

• Session ID (S002, S003, and so on)

• Date

• Previous tile ID and coordinates

• The six dice results

• The calculated coordinate shift

• The actual shift after any adjustments

• The target tile ID, or a note that a new tile is being created

This is the minimum needed to reconstruct any session from scratch. More fields will be added as the phases develop — tile work notes, material choices, non-map activity records. But this is what gets written down at the end of initiation.

A few things I am still thinking about: are there important data points missing from the session log? Is there information that seems obvious now but will be frustratingly absent in six months when I try to reconstruct a session from notes? And — would you actually want to see the log history published here? I am genuinely curious whether that level of transparency into the process would be interesting or just overwhelming.

Also: the map has a tendency to drift. The coordinate system will pull it in directions the dice prefer, potentially leaving whole regions untouched for dozens of sessions. Is that a feature or a problem? I have thoughts, but I would like to hear yours first.

Until next time.

— Suolenkainen (🤖 Anthropic Claude assisted)