jonnifyagile agents

A7 · orientation dive 1

Why an end-to-end exemplar

A6 makes each rung production-grade. A7 exists because a method is proven only when the whole loop runs end to end on a real system — a sequence of hardened rungs is not a demonstrated, composed run.

The failure A7 prevents is the one every teaching course risks: techniques proven in isolation that never compose into a shipped system. Seven correct lessons do not add up to a deployed Portal; one composed run does. Select a mode on the right to watch the same seven moves either stay separate lessons or compose into one run that ships.

Isolated moves vs one composed run · pick a mode

seven moves · the same techniques, run two ways A1why A2decompose A3roadmap A4spec A5brief A6harden A7run VERDICT no shipped system seven separate lessons, each proven on its own fragment — never assembled a method is proven by one whole run, not seven separate lessons moves composed: 0 of 7 · ships: no
Taught in isolation — 7 separate lessons, each proven on its own fragment; no composed run, so no shipped system. The moves are correct but never assembled.

Why the run is its own Part

By the end of A6, every move the course teaches has been proven: A1 made the case for thin slices, A2 decomposed a vision, A3 planned delivery, A4 defined and proved a rung, A5 briefed an agent to build it, and A6 hardened the result to production quality. Each was demonstrated on a worked fragment. None of that, on its own, is a shipped system.

A7 is its own Part because composition is a separate claim from correctness. A technique can be correct in isolation and still never assemble into a deployed product — the seams between moves are where real systems fail, and seams only appear when the moves run in sequence on one codebase. The exemplar makes that claim concrete: it runs decompose, roadmap, spec, brief, build, harden, and accept once, uninterrupted, from an empty repository to a deployed Portal. The proof is the whole run, not the sum of the lessons.

what A6 leaves
Each rung made production-grade in isolation — correct under failure, but not yet assembled into one shipped run.
what A7 adds
The composition: the moves run in sequence on one codebase, from nothing to production, with the seams exercised.
the failure prevented
Techniques taught in isolation that never compose — correct lessons, no shipped system.
the proof
One whole run that ships, not seven separate demonstrations that each pass on their own fragment.

This dive answers the chapter's reason-to-exist: A6 hardens each rung; A7 exists to prove the hardened rungs compose into a shipped system — a claim no single technique makes on its own.

The A6→A7 seam

Read the chain backward and the seam is exact. A6 delivers hardened rungs. A7's job is the one thing A6 does not produce: a run that assembles those rungs into a shipped system. Selecting a state below reports what each layer holds — what exists after A6, what is still missing, and what the composed run produces.

The seam · pick a state

read the chain backward · what each layer holds A6 DELIVERS hardened rungs — each production-grade in isolation A7 ADDS a composed run — the rungs assembled in sequence on one codebase THE OUTPUT a shipped system — the deployed Portal, zero to production the seam: hardened rungs exist; the composed run does not, until A7
Hardened rungs — what A6 delivers: each rung made production-grade in isolation. This is the input A7 composes; on its own it is not a shipped system.

The seam is the load-bearing reason A7 follows A6: hardened rungs are the input; the composed run that ships them is the work A7 does.

On the Portal

The Portal is what makes the claim testable. It is one real, multi-surface system — a branded store, an event-sourced engine behind one facade, a Phoenix web app, a Telegram bot, and a student dashboard — carried from an empty repository to a deployed platform. Running the loop on the Portal means the seams between the moves are exercised on a system that actually has to run, not on seven disconnected fragments.

The companion Functional Programming in Elixir course builds the Portal's surfaces for real, with the OTP internals it cites rather than re-teaches; its F6 Phoenix chapter is the web surface, shipped end to end from a real roadmap. A7 is the agile-side mirror: the same system, run as one composed loop, to show the method holds whole.

The principle

A sequence of correct techniques is not a demonstrated method; the method is the whole loop run once, from nothing to production, with the seams between moves exercised.

On the Portal

A7 runs decompose → roadmap → spec → brief → build → harden → accept once on the Portal, shipping the five-surface platform from an empty repository to production — one run, not seven lessons.

The exemplar closes the chain by running the loop end to end, not by adding another technique.

Recap — the proof is the whole run

A7 is its own Part because composition is a separate claim from correctness. A6 leaves every rung production-grade in isolation; A7 assembles them into one uninterrupted run on the Portal, from an empty repository to a deployed platform. Seven correct lessons do not add up to a shipped system; one composed run does — and the seams between moves, where real systems fail, only appear when the run is whole. The next dive names what that run covers: the seven steps A7.01–A7.07, the zero-to-production arc.

Next in A7: What it covers — the seven steps A7.01–A7.07 as the zero-to-production arc, grouped into three milestones, with the per-step detail deferred to the chapter's triad.

References

Primary sources for running the whole system end to end, the inspect-and-adapt loop run whole, and keeping the result releasable — and where the moves A7 composes are taught in the course.

Sources

Related in this course