influences.sgit.ai — the influence register of Dinis Cruz
An inspiration is something that moved you. An influence is something that shaped the work — and these are 25 of Dinis Cruz's ↗: the people, works, topics and things behind twenty-five years of his software, with the idea worth taking from each and the exact place in his code where you can check that it landed.
Twenty-five years in security and software · former CISO · former OWASP board member and Summit organiser · founder of sgit.ai↗. More about him at open-source.sgit.ai/about↗ and docs.diniscruz.ai/about ↗, and why this register is his to make ↗.
Source: https://influences.sgit.ai/index.html↗ · site v0.4.0 · markdown twin of the front page.
Eight ideas worth stealing
Each one distilled to a single sentence you can apply without having read the book or watched the talk — which is the test the format sets. Every entry behind them names where the idea shows up in real code, at which version, and where it is still missing.
- Bret Victor (traced) — Immediate connection between creator and creation: a creator needs to see what they are making, as they make it. The entry↗
- Tim Berners-Lee & the Semantic Web (traced) — Meaning should be machine-readable, so that independent parties can exchange it without agreeing on a schema first. The entry↗
- Design, with a capital D (traced, discovered) — Design is not decoration applied after engineering: it is how the thing works. Start from what the person is trying to do and work backwards to the simplest interaction that does it. The entry↗
- Flow, and coding in the zone (traced) — Clear goals, immediate feedback and a challenge matched to skill produce the zone, so a methodology's job is to protect those three conditions rather than to optimise throughput. The entry↗
- Niklas Luhmann & the Zettelkasten (traced, discovered) — External memory works as a thinking partner when its units are atomic, uniquely addressed, densely linked and allowed to grow structure rather than being filed into one. The entry↗
- Karl Popper & falsifiability (traced) — A claim earns its status by being refutable. If nothing could show it to be wrong, it is not saying anything. The entry↗
- Christopher Alexander (discovered) — Code is a space that people inhabit, and a language of patterns is what lets a place be built by many hands and still be coherent. The entry↗
- Simon Wardley & Wardley Maps (traced) — Situational awareness before strategy: draw the map before you argue about the move. The entry↗
These are eight of 25 ↗. The rest run from a Canadian drummer to a 1945 essay about a machine nobody built.
Five things a register like this is not supposed to contain
- An influence that is disagreed with. The Semantic Web entry credits the ambition and then says exactly where the community went wrong — meaning attached to nodes rather than derived from edges. Two rows of its trace table record the influence being inverted rather than implemented. Tim Berners-Lee & the Semantic Web↗
- An influence that became a mandatory step in a pipeline. Somebody read that good design is invisible — that you only notice it by going back to the previous version and feeling the loss — and turned it into two questions every interface change has to answer: is it simpler? would reverting feel worse? Not a quote on a wall. A gate that can fail. Design, with a capital D↗
- Influences he never claimed. Seven entries were found by searching his work rather than by asking him. They are published as claims about him, and he is asked in public to strike the ones that do not belong. The discovered tier ↗
- Entries with no evidence at all — published anyway. Seven influences are on Dinis Cruz's own list and the search found nothing. One of them has no public record and never will. Music, and playing in a band↗
- Rows that say absent. Every entry lists the patterns from its anchor work that the codebase does not implement, written precisely enough to pick up as a work item. The gaps, as build specs ↗
How to read the register
Twenty-five entries, sorted by how well the claim is evidenced rather than by how much they mattered. The tier is a statement about evidence, not importance — a stated influence may well be the deepest one on the list.
- TRACED — 15. His own writing names them. The entry lists the files and the claim can be checked against them, which is all traced means. The definition ↗
- STATED — 7. He says so and the written record is silent. Published as stubs with the research plan visible and a hypothesis clearly labelled as one. The definition ↗
- DISCOVERED — 3. His work says so and he never did. Claims about him, awaiting his confirmation or his correction. The definition ↗
Then: all 25 entries ↗ · the influence map ↗, computed from the register rather than drawn · seven blocks per entry ↗ · the wider library ↗, a list of addresses rather than a collection.
Articles
- An influence is a claim you can check↗ (2026-08-26) — an introduction to this register: the one move that separates it from a reading list, the five ideas worth stealing from it, the three tiers and what moves an entry between them, and the four things the site cannot do.
All articles ↗. Each is a markdown file this site renders, so take the .md
if you want to republish it.
One thing to know before you believe any of it
A trace table cannot show causation. It shows that a pattern from a talk appears in a codebase. It cannot show that the talk put it there — and the finding that started this whole register is exactly that distinction: half the estate's strongest features turned out to be unknowing implementations of a talk's demos, built on instinct years later by people who had not connected the two. The honest claim is that the instinct and the principle agree. It is weaker than this shaped the work, and it is the one the evidence supports.
There is also an obvious conflict of interest, and it is worth naming on the way in: this is a register about Dinis Cruz, published by the project he founded, claiming his own work was shaped by a list of well-regarded thinkers. Association is flattery, and it costs nothing to claim. The defence is the only one available — every claim here is checkable by somebody who does not share the incentive.
All site content CC BY 4.0 — Dinis Cruz, with AI co-authorship (Claude, Anthropic). The anchor works belong to their authors and are linked, never rehosted.