The first chapter of the Start here track and the first chapter of Phase D,
rewritten in the Eve edition on 10 September 2026. It follows the plan in
EVE-V1-REDESIGN-TODOS.md and the track brief in
start-here-track.md. The receipt is under
verification/.
The job#
This is the front door. Ten products share one platform, and the fastest way to get lost is to read a V number as a version, a release as a product, or a shared capability as somebody's feature.
Running example. A mystery clue arriving quietly on the Rail: V10 decides whether to show it, V8 proved the case is solvable, and the investigation happens somewhere else again.
Three learning promises (the cover)#
- Product, release, domain, foundation — four words that answer four different questions, and the mistakes each one prevents.
- What each product is for — ten experiences in order, each with the promise that makes it a separate product rather than a mode.
- How to follow one thing across two — experience, authority, composition and evidence.
Three things to remember (the close)#
- A V number names a product. Not a version.
- Shared does not mean unowned. A domain has one owner and many users; a foundation has none and is composed by all of them.
- Four questions, in order. Experience, authority, composition, evidence.
Bridge to chapter two. Which track to read, in which order, and what each one assumes you have already read.
Sections#
Three sections, opened by a section-divider that names all three.
| Section | Slides |
|---|---|
| One platform, ten products | the map, the four words, V1, V2, V3 |
| Worlds, cells and realms | V4, V5, V6, V7 |
| Proof, learning and attention | V8, V9, V10, and the method for following a capability across owners |
Treatment decisions#
- Two legacy
graphdiagrams retired.portfolio-productsandportfolio-v3-embodimentwere bothdiagrams.jsonoverrides; rewriting them as real graph diagrams leaves sixteen. - The glossary is a slide, not an appendix. D0 asks for a
glossary-cardsslide in each track's chapter one, and here it earns its place: the four words are the chapter's first promise, and each term is paired with the specific misreading it prevents rather than a definition alone. - Ten products, ten different treatments where the product justified one. Seven of the ten were planned as card grids, which would have been seven identical slides in a row. V2 became a compare panel because its promise is literally two-sided; V3 a graph diagram because three tenants share one set of services; V7 a state machine because a realm has a lifecycle with an admission gate; V8 and V9 step journeys; V10 a layer stack because the Rail's whole job is what may cross between attention bands.
- No capture. Nothing in
assets/shows a product surface from V2 to V10, so the treatment checkbox requiring one real capture stays open.
What the source review needed#
authoring/orientation-source-review.json binds every slide of this guide to
the exact pinned spans its claims rest on. The rewrite moved twelve fingerprints
and added six slides, so six new claim-level reviews were written and the
checker's scope counts moved from 33 instances / 32 unique slides to 39 / 38.
The chapter cover now also cites the V10 and V8 sources, because it is where the
running clue is introduced and a claim should be cited where it is made.