influences.sgit.ai / shipped

Tier movement — the changelog

This site has two changelogs. One is the release history, which says what shipped. This one is the interesting one: what changed about the claims — an entry moving tier because a briefing document arrived or Dinis Cruz confirmed it, a discovered entry struck out because it turned out to be a citation, or a trace row flipping absent to implemented because the estate built one of the gaps an entry specified.

Why this page exists at all. A provenance site whose provenance changed silently would be self-refuting. If an influence quietly moves from Dinis Cruz said so to and here is the evidence, that is the single most informative event this site can record, and burying it in a diff would waste it.

Movements

DateEntryMovementWhat caused it
25 Aug 2026 Design, with a capital D discovered → traced Dinis Cruz confirmed it in writing, the day the commissioning pack shipped and before this site existed. The influence had been surfaced by mining a role definition and two briefs, and the trace was already deep — including the Jonathan Ive test as a mandatory validator for every UI change. The first DISCOVERED → confirmed transition on record, and the model for the rest.
25 Aug 2026 Luhmann, Csíkszentmihályi, the design canon discovered → traced Promoted at mining time on the strength of the evidence rather than by confirmation — each had a dedicated article or a load-bearing citation behind it. They keep a discovered → traced badge, because how an entry got here is part of what it is, and their confirmation is still open as Q2.

What is expected to move next

Not a prediction — a list of the transitions that are currently blocked on something specific, so that when one happens it is visible that it was waiting.

EntryWould moveBlocked on
Christopher Alexanderdiscovered → traced or removedQ2 — the citation is load-bearing, which is an argument and not a confirmation
Vannevar Bushdiscovered → traced or removedQ2 — one citation deep, and reached through another entry
Team Topologies & Cynefindiscovered → traced or removedQ2 — a single mention in a 2019 hiring post
The 7 stated entriesstated → tracedA briefing document each — N1 through N7. Two of them, Peart and the band years, are the ones nobody else can research.
Bret Victor's trace tableempty → rowsQ1 — the full register text is with Dinis Cruz. Entry one is not complete until it arrives.
Seven entries with pending trace tablespending → rowsR1 — the corpus is cited rather than resolved, because this repository holds the website and not the estate.

Gaps the estate could close

The other kind of movement. Every full entry publishes its gaps as build specs, and a gap that gets built flips a trace row from absent to implemented. That turns an influence page into something unusual: a backlog that was written by reading a talk.

None have been closed yet — this is the first release with a register. The ones most worth doing first, on the site's own reading:

What would be recorded here if this site were wrong

The transitions above are all additive — entries getting stronger. That is what you would expect from a register whose first release is also its first content release, and it is also exactly what you would expect from a site that only ever confirms itself.

So, precisely: a DISCOVERED → removed row would appear here if Dinis Cruz said an entry was a citation rather than an influence, naming what was struck and why. A retraction row would appear if a trace row turned out to be wrong. A counter-influence entry would appear in the register if there is something everyone assumes shaped the work and did not. None of those has happened, and until one does, this page is evidence of ambition rather than of practice. The disclosure page says the same thing at more length.