See the reports a party runs prototype
Table of Contents
| Item | Where |
|---|---|
| Prototype | the prototype |
| Gemini review | none |
| Group | Reporting prototypes |
1. Summary
This prototype argues what the front door of reporting is, before any of it is
built. It carries three structural variants of the same screen, switched with
?variant, and it answers to
See the reports a party runs.
It is throwaway. It is not production code, it reads no service, and every value on it is mock data.
2. Detail
2.1. The question
A party runs a number of reports, some on a schedule and some on demand, and each of them has a latest run that either went well, went wrong, or never happened. The open question is what a person lands on when they open reporting. The journey can answer that five ways and still be true to itself, which is why this is a prototype and not a specification.
2.2. The three variants
Switch with the bottom bar, or with ?variant.
| Variant | Name | The structural argument |
|---|---|---|
A |
Index-first | The catalogue is the page: one row per report, its schedule, its owner, and the health of its latest run. A run is a detail reached from a report. |
B |
Ledger-first | One dense table of runs across the party, newest first, report as a column. The catalogue is a filter, not a page. |
C |
Attention-first | Only what needs a human: failed runs and runs that completed carrying errors. Clean runs exist behind a filter, off by default. |
The variants disagree about structure, not colour. A starts from the thing that is scheduled. B starts from what happened. C starts from what is wrong, and is the only one that hides work rather than ranking it.
2.3. The health words, and the thing a reader gets wrong
The latest run of a definition is one of five: failed, completed with errors,
running, never run, completed clean. The summary tallies them, and each row
carries its error count split into errors and warnings.
The point the page exists to make is that a completed run is not necessarily clean. A row that completed while carrying errors is drawn apart from a clean one without opening anything. A definition that has never run is not clean either: nothing ran and nothing was checked.
2.4. What is fake
Everything. The definitions, parties, schedules, owners, run times and error counts are mock. Nothing reads a server, and the application has no reporting UI today. The page says both, at the top, in the strip and the note.
2.5. Where the code lives
doc/prototypes/reporting/see-the-reports-a-party-runs/. Three files, plain
HTML, CSS and JavaScript, no framework and no build step. The site build finds
the directory and deploys it where it sits.
2.6. The verdict
None yet. This page is on the branch that proposes it and waits for review. Whichever variant is chosen, the health tally and the split error count stay: they are the part the other four prototypes depend on.
2.7. What it does not settle
Whether the five journeys become one area, one screen or two. This prototype argues the front door only, and the area behind it is judged from the other four.
3. See also
- Reporting prototypes — the group, and the axis this prototype turns on.
- Reporting journeys — the five journeys these answer to.