Bret Victor
The model entry, and the reason this site has a format at all. The register for Inventing on Principle was written as a working document before anyone asked what a site made of such documents should be called.
Block 1 The anchor
Inventing on Principle — Bret Victor — CUSEC · 2012
↗ the anchor work ↗ transcript
Linked, never rehosted. This site explains why the work resonated and traces where it was applied; the work itself stays where its author put it.
One anchor per entry: not the whole career, the single work that did the shaping. Victor's wider work is in Block 7.
Block 2 In his own words
First person is the house style on this site and nowhere else in the network — what resonates with me and why is the genre, and paraphrasing it into corporate third person would destroy the evidence. Until the briefing document for this entry arrives, this block carries the first-person statements that already exist in the corpus, cited by file.
That is the kind of experience and feedback loops I'm trying to create here.
The line the register opens on, and the reason it exists. The register is a working document about the estate's viewer, not an essay about Victor.
the problem with dogma is not that it exists, but that you stop asking the question.
From a section titled The Brett Victor Problem — in an article about education policy, not developer tooling. One influence applied in two unrelated domains, which is itself trace evidence: an idea that only shows up where you first met it has not been absorbed.
Half the viewer's strongest features were already unknowing implementations of its demos, which is evidence the instinct and the principle agree; this register makes the agreement deliberate, so the next agent extends it on purpose rather than by luck.
The closing argument, and the standard every trace table on this site is aiming at. Note how careful it is: the instinct and the principle agree is a much weaker claim than the talk caused this, and it is the one the evidence actually supports.
The talk's argument is not about tools. It is that ideas can only grow where their creator can see them growing, and that most of the ways we build software put a compile step, a deploy step or a refresh between a person and the thing they are making. Every one of those steps is a place an idea dies.
What makes this the founding entry is what the register found when it was written: the estate's viewer had already implemented several of the talk's demos without knowing it. That is a weaker claim than Victor caused this and a more interesting one — the instinct and the principle agree, and writing the register is what turns an agreement into something the next agent can extend deliberately rather than rediscover.
Block 3 The principle
Immediate connection between creator and creation — a creator needs to see what they are making, as they make it.
The test the format sets for a principle is that it must be usable by someone who has never consumed the anchor work. This one passes: you can apply it to a build pipeline, a form, a CLI or a lesson plan without having watched the talk.
Block 4 The trace table
No trace table yet. The table exists and this site does not have it. The register's own trace table — five patterns from the talk, each with where the repository stands and what is missing — was written by Dinis Cruz and supplied in conversation; it was not retained across a context boundary and this environment cannot reconstruct it. Inventing rows would be exactly the failure this format exists to prevent, so the entry ships with the table empty and the gaps it does name published in Block 5. This is the first ask in the briefing queue — Q1 — and entry one is not complete until it is answered.
Block 5 The gaps, as build specs
Patterns from the anchor the estate does not implement yet, written precisely enough that an agent could pick one up as a work item. This is what makes an influence page forward-looking: it is provenance and backlog in the same document.
View time travel
The register names this as a specified gap: the viewer cannot step backwards through the states a view has been in. Victor's demos treat time as a dimension you can scrub, not a sequence you can only move forward through — the binary-search-for-the-bug case in the talk is the canonical demonstration. Precise scope is in the register text this site does not yet hold.
Relayout ghosts
The register's second named gap: when a layout changes, the previous positions leave no trace, so a person cannot see what moved. Victor's argument is that a change you cannot see the shape of is a change you cannot reason about. Precise scope is in the register text this site does not yet hold.
Block 6 The checklist
What to ask of new work in this influence's light. The checklist is the influence made operational — the part an agent can run without having consumed the anchor work.
- Does the person see the effect of their change immediately, or is there a step between them and it?
- If there is a step, is it there because it has to be, or because it was easiest to build that way?
- Can they see the state the system is in, or only the output it produced?
- When something goes wrong, can they scrub backwards to the moment it went wrong — or only re-run it and hope?
- Is a value on screen something they can reach into and change, or only something they can read?
- Would someone who has never used this be able to tell what it does by moving one thing?
Block 7 The wider library
The rest of the work, linked and never rehosted, each item with one line on what it adds. The full union of every entry's Block 7 is at /library/.
- The Future of Programming — the 1973 costume talk — a history of ideas that were abandoned rather than disproved ↗
- Media for Thinking the Unthinkable — why the representations we use bound the thoughts we can have ↗
- Up and Down the Ladder of Abstraction — the interactive essay about moving between the concrete and the general — arguably the most directly applicable to graph work ↗
- Stop Drawing Dead Fish — on tools that produce artefacts versus tools that produce behaviour ↗
- Drawing Dynamic Visualizations — direct manipulation applied to data visualisation itself ↗
- The Humane Representation of Thought — the widest version of the argument: thinking with the whole body, not a rectangle ↗
- Dynamicland — the research programme the ideas went into ↗
The corpus evidence
What the mining run found, by path. These are the files that put this entry in its tier — the claim on this page is checkable against them, which is the whole point of the format.
| Path in the corpus | What it carries |
|---|---|
sources/register__bret-victor__inventing-on-principle.md | Dinis Cruz's immediate-connection register — structural summary published at /documents/victor-register.html; the full text is still with him |
docs.diniscruz.ai/docs/2025/07/04/the-joy-of-programming-in-the-age-of-ai-assisted-development.md | builds its central argument on Victor's immediate-feedback thesis, and quotes the talk's core line about creators needing an immediate connection to what they create |
SGraph-AI__App__Send/team/humans/dinis_cruz/briefs/03/28/v0.19.7__article__education-gaps-git-opensource.md | a section titled The Brett Victor Problem, applying the talk's point about dogma to education policy |
Paths are as recorded by the mining run behind the commissioning pack (v0.33.62, 25 August 2026). This repository holds the website, not the corpus, so they are cited rather than resolved — R2 in the comms queue.
Where this sits
No nesting and no sibling site: this entry stands on its own. The influence map draws every relation the register records.