Thing Pong - My Zettelkasten Has A Ridiculous Name And That's The Point
08-03-2026
I have a problem that most productivity systems are not designed for. I don't have one area of focus. I have fighting games, generative AI, sculpting, Go, music theory, game design, programming philosophy, and about fourteen other things I am convinced I will pursue seriously. Simultaneously. At all times.
The classic solution is folders. Give each interest its own folder, keep things tidy, revisit occasionally. The problem with this is that the insight that makes fighting game neutral theory feel structurally identical to functional programming architecture does not live in any folder. It lives in the gap between them. And folders don't have gaps. Folders have walls.
So I found Zettelkasten. And like most people who find Zettelkasten, I went a little bit insane for a while.
Zettelkasten is a note-taking philosophy invented by Niklas Luhmann, a German sociologist who produced over 70 books and 400 articles in his lifetime, and when asked how, pointed to a wooden cabinet of index cards. Around 90,000 of them. One idea per card. Each card linking to others. He called it his thinking partner. Not his storage system — his partner. The cards thought with him.
The idea sounds simple. Keep your notes atomic — one idea per note — and link them to each other. The complexity you get for free, emergent from the connections accumulating over time. It is, in a sense, the functional programming of note-taking. Small pure units, composed into something larger than what any individual unit contains.
I read all of this and thought: yes. And then I tried to implement it and got stuck on the most fundamental question possible. What is a note?
Most Zettelkasten guides tell you to categorize. You have literature notes, fleeting notes, permanent notes. Some people add project notes, reference notes, hub notes. And suddenly you have a second organizational system, this time made of note types instead of folders, and the same walls as before, just renamed.
My solution, which took longer to arrive at than I would like to admit, is embarrassingly simple. Everything is a Thing.
Everything is a Thing
A concept is a Thing. A place is a Thing. A person is a Thing. A tool, a moment, a question, a half-formed opinion — all Things. There is no taxonomy. There is no hierarchy. There is ontological flatness and nothing else. The moment you name something and link to it, it becomes real and traversable in the graph. Naming is an act of ontology. You are not cataloguing reality. You are constructing it, one link at a time.
Thing notes carry no topic tags. Just a maturity state. Forcing a taxonomy onto them creates friction and false categorization. Fighting games and music theory don't belong in separate cabinets. They belong in the same graph, where the note that connects them can surface naturally when you write something that genuinely touches both.
This is why I called the system Thing Pong. Yes, it is a pun. Yes, it is intentional. If I am going to maintain a second brain for the rest of my life, it had better reflect my personality, and my personality is not compatible with names like "Personal Knowledge Management System". Thing Pong is a pun on Ping Pong — not the sport, but the metaphor for the back-and-forth of ideas during brainstorming. The notes bounce between each other. The graph is the table.
Notes Have Types, But Not the Types You're Thinking
Even with ontological flatness at the Thing level, notes themselves have types — and this turns out to matter a lot, but for a different reason than you'd expect. The types are not categories. They are arcs.
An idea note matures from fleeting to shareable. A question note matures from stub to answered. A tinker note matures from in-progress to polished. A source note matures from bookmark to genuine reference. A diary note is a time-indexed field report and doesn't mature at all — it just accumulates as a record of where your head was on a given day.
The types tell you what to expect from a note, not where it belongs. And they carry different atomicity expectations. An idea note holds exactly one claim, no more. A source note can be long — length reflects the material, not a failure of discipline. A diary entry can ramble, because the point is capture, not precision.
What unifies them is the linking behavior. Every type links outward toward Things and ideas and questions. Only one pairing is genuinely bidirectional: idea and question. They are allowed to point at each other because the relationship between a thought and the curiosity it generates — or the curiosity that generated it — doesn't have a single natural direction. Whichever note existed first holds the link. The other collects the backlink.
A Note Matures in Two Dimensions
Here is the part that most Zettelkasten implementations get wrong, including my earlier versions: maturity is not a single axis.
A note can be beautifully written — precise, complete, a single clean claim — and still be effectively dead. If nothing links to it, no traversal will ever reach it. It exists in the vault the way a word exists in a language nobody speaks. Technically present. Practically unreachable.
Maturity in Thing Pong is compound. It requires two axes advancing together.
The first is Depth — the note captures the minimum complete idea. It is a standalone atom. Remove a sentence: did the note lose its claim, or just a restatement of something another note already says? If the latter, cut and link. The test is ruthless and the result is notes that cannot be made smaller without losing their point.
The second is Breadth — the note cannot be removed from the graph without leaving a hole. It has live outgoing links and inbound links that make it reachable. A note with no inbound links will never appear in any blog post traversal, regardless of how well-written it is. Depth without Breadth is a beautiful dead end.
The maturity states follow from this. A raw note is a fragment or stub, no meaningful links. Warm means coherent atom, at least one live link in either direction. Hot means dense, actively referenced, alive. Cooked means complete, graph position stable. Digested means second nature — structurally load-bearing, the kind of note you find yourself linking to from everywhere because the idea has become part of how you think.
Links Are Not Filing Decisions
Now here is the part that changed how I actually write.
I used to treat linking as an organizational decision. You link a note to where it belongs, like filing a document. The link is a classification. This is wrong, and it is wrong in a way that quietly ruins the whole system without you noticing, because everything still looks tidy.
A link is not a navigation decision. It is a sculpting gesture. When you are writing a note and a word surfaces that has more to it than this one sentence — wrap it. [[Like this]]. You are cutting an injection point into the surface of the prose. You are not deciding where the thought goes. You are marking that this concept has depth.
The mantra: Write where you are. Wrap what has weight. The graph builds itself.
The origin context holds the link. The subject collects the backlink. You never have to file anything, because there is no wrong note to be in. Capture and canonicalization are the same act. The moment of writing decides the link direction automatically — where the thought was born is the context, and context is causality.
There is a consequence of this that I find genuinely beautiful. The things you wrap in links are the things you want to revisit. Everything else is conversational filler. Linking is not just graph construction — it is an act of curation. Looking back at a note's links is like reading a record of what you actually cared about when you wrote it, not just what you intended to care about.
A vague word attracts vague connections, by the way. A specific name — like quiddity rather than "essence" — accumulates only the notes that genuinely mean it. The label is itself an abstraction, and the more specific the abstraction, the more precisely the graph clusters around it. Name Things carefully.
Even Empty Notes Are Doing Something
The consequence of this model that surprised me most: a stub with no content still participates fully.
A Thing note can be nothing but a name — no body, no claims, no links of its own. And it still functions as a gravity point. When two notes written months apart independently reach for the same word, the stub holds them together without knowing it. Open it and find a backlink timeline — every moment that concept was alive in your thinking, in chronological order.
It is not waiting to be filled. It is already working. The empty Thing is not a placeholder. It is a record of convergence.
This is what I mean when I say the graph organizes itself. You don't architect it. You write, you wrap, you create stubs freely and aggressively whenever a concept surfaces — even half-formed, especially half-formed — and the structure emerges from the accumulated trace of your attention.
The Output: A Post Is a Path, Not a Document
And this is where the system becomes genuinely strange and, for me, genuinely exciting.
A blog post does not exist anywhere in my vault. It is not stored in any note. It emerges from a traversal — a path chosen through the graph, a sequence of nodes visited, connective tissue supplied in the writing. Two posts can start from the same note and arrive somewhere completely different depending on which links get followed. The graph is generative. Notes are vocabulary. The post is a sentence you construct from them.
The only requirements are that notes are atomic — so they contribute one clean thing to the path — and reachable — so the traversal can actually find them. A note with Depth but no Breadth is a word in a language only you speak, in a room with no doors.
This also means the Zettelkasten does something that offloading notes to a folder system never does: it sharpens thinking in the act of capture. The requirement of atomicity forces you to distinguish ideas from each other precisely. You cannot be vague about where one thought ends and another begins, because the vault will accumulate the ambiguity and reflect it back at you as noise. The discipline of one claim per note is not a constraint on expression. It is a clarification of thought.
This article is that. The nodes were already there — requirements, ontology, note types, maturity, linking, stubs, traversal. I walked them. The argument assembled itself from the path.
Which is, come to think of it, exactly what Luhmann was describing with his cabinet. Not storage. A conversation partner. Something that thinks back.
-- Written with Claude using the Obsidian MCP to link original ideas from my Zettelkasten
Nodes visited:
My Requirements for a Zettelkasten— the problem, the entry woundZettelkasten+Atomicity of Zettel— Luhmann, the philosophyEverything is a Thing+Things are unlabeled in Thing Pong— ontological flatnessName for my Zettelkasten System - Thing Pong— the name, the punNote Types in Thing Pong+Atomicity applies differently per note type+Link Direction by Note Type— the type system as arcsDepth of Zettelkasten Notes+Breadth of Zettelkasten Notes+Maturity States in Thing Pong— compound maturityLinks as Sculpting Gestures+Capture vs Canonicalization+Links reveal what you care about— the linking modelNaming precision determines what accumulates on a Thing— specificityEmpty Things still participate in traversal+Aggressive Note Creation in Thing Pong— stubs as gravity pointsA blog post is a traversal of the graph+Benefit of my Zettelkasten— the output model and the meta-close