Format drift, quantified
Across the 17 Explorer role directories, the schema in the ROLE.md format is stable. The presentation is not — and the migration between the two markup dialects is what lost the format's best property.
| State | Count | Roles |
|---|---|---|
Identity as table (| **Field** |) | 7 | architect, conductor, designer, dev, devops, dpo, qa |
Identity as bullet list (- **Field:**) | 6 | appsec, cartographer, grc, historian, journalist, librarian |
No ROLE.md at all | 4 | advocate, alchemist, ambassador, sherpa |
Two markup dialects of one schema, plus four directories that exist without a
definition. The schema is stable; the presentation is not. A parser reading these files
must handle both — which is itself an argument for publishing a canonical
machine-readable form, which is exactly what
data/roster.json (the pack's
teams__roster.json) is: the first cut at one.
The regression, restated plainly. The bullet-list dialect is the older one, and its Central Claims are written as falsifiable failure conditions. The table dialect is newer, better formatted, easier to scan — and its Central Claims are descriptive, which means they cannot be checked. A format migration silently dropped a semantic constraint, and nothing in the estate's own history flags that this happened. It took an external measurement to notice.
The four empty directories, and one more
advocate, alchemist, ambassador and sherpa have a directory in Explorer's
team/roles/ and no ROLE.md inside it. Town Planner's librarian is
the fifth: a directory with no definition, in a different team.
07__ of the pack is explicit about
what happens to these on this site: they are published as gaps, and this site does not
write a definition for them. An invented role definition would make every real one on
this site unverifiable — the whole format's claim to be checkable rests on never doing
that. See each one's page in the roster for exactly
what state it is in, per team.
Why this is worth a page on its own
Explorer/Villager may have been driven by context limits as much as by deliberate design — running one enormous team is impractical, and splitting it is convenient. The same honesty applies here: a format that drifted without anyone deciding to change it is not a scandal, it is what happens to conventions under real production pressure. The useful move is measuring it and restoring the property that mattered, not pretending it never drifted.