Skip to content
Draft
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Original file line number Diff line number Diff line change
Expand Up @@ -8443,3 +8443,119 @@ report Metaschema constraint findings the JSON schema does not express:
step, and prop names outside the allowed set. These are the publisher's
modelling choices, they do not affect the verifier's checks, and they are left
as they are.

## Session: the IBM pairs arrive as single files

### Two combined component definitions added

**Date:** 2026-10-09
**Scope:** examples/component-first, the example index and lock, the OSCAL
inventory, one pinned test
**Result:** `--all --offline` 2701/2701; unit tests 68/68.

**What was added.** Vikas Agarwal supplied two files on issue 16, each holding a
service component and the validation component that checks it:
`idservice-oscap-validation-cd.json` (IDSERVICE with OSCAP) and
`cos-ansible-validation-cd.json` (COS with Ansible). Every component in them is
identical to its counterpart in the four separate files. The separate files stay,
because snippets and verifier checks resolve against them; the pairs are an
easier way into the same content, not a replacement.

**What was changed in them.** Only the document `uuid` and `last-modified`. All
four separate files share the document UUID `bbdf7ee6-…`, and the combined files
arrived with it too, so each combined file now has its own:
`5387e03f-f015-4c6c-a2b8-384f2319c087` for IDSERVICE with OSCAP and
`51f93843-4b96-4794-a590-1cbc6bad9b73` for COS with Ansible. Nothing inside the
components was touched. The UUID shared by the four separate files, and the
control-implementation and implemented-requirement UUIDs shared by the COS and
IDSERVICE components, are left as published.

**Raised, not changed.** The three OSCAP checks give
`target-component-uuid` `0c410eb1-…-c80126a984f9`, which names no component. The
IDSERVICE component is `…984e9`, and the IBM assessment plan uses `…984e9`. The
same value is in the separate `oscap-validation-component-definition.json`. It
was put back to the author on issue 16 rather than corrected here.

**What moved with them.** The IBM set is 8 files, so `tools/source-lock.json` and
the pin in `test_source_inputs.py` say 8, and the index was rebaselined.
`oscal-artifacts.json` lists 6 component definitions for component-first, up
from 4, since it counts files. `ibm_cdef_files` in the corpus figures stays at 4:
its glob is `*component-definition.json`, which the combined `-cd.json` names do
not match, so the same components are not counted twice.

## Session: issue 16, the review closed out

### What the issue asked for that had not reached the site

**Date:** 2026-10-09
**Scope:** data/questions.json, data/stakeholders.json, data/six-questions.json,
component-first.html, two stakeholder checks and the stable question ids
**Result:** `--all --offline` 2708/2708; unit tests 68/68; approach pages within
budget at a 9.5 per cent spread.

The review on issue 16 was read end to end against the published pages. Most of
the accepted edits were already in. What follows is what was not.

**Two accepted moves had become deletions.** The reply on the issue accepted
moving the observation-to-asset question and the desired-state question to the
shared section. Both were deleted instead. Both are back, in the shared section,
under their old ids (`proposal-4`, `paper-comp-2`), and the ids are now held by
the stable-id check. The observation question is reframed as the issue asked:
it concerns the current assessment-results model, an observation can name its
producer through origin actors but not the check it ran, and the component-first
examples propose the asset UUID and check identifier on each observation.

**The component-first party table now reads as the approach describes itself.**
The opening post rewrote five rows. Fitted to the table's two-sentence notes and
to the page budget:

- The mapping provider publishes framework-to-framework mappings, a mapping
collection of its own, and does not tie controls to product rules.
- The technology provider's rules carry the framework tie and the evidence type
that proves each rule.
- The policy engine provider's checks carry no framework tie.
- The system owner also selects the validation components the plan uses.
- The auditor's run makes observations and findings; observations roll up to
control-level findings through check, rule and control.

This retires the "software or validation" marker on the mapping provider's row,
which had stood for the open question of which component definition carries the
tie. The approach's answer is the rule component. Two stakeholder checks move
with it: a mapping collection may now be the mapping provider's on two
approaches, and on both it is its own document. The marker's own rule is kept,
unused, for any approach that leaves the question open again.

**`which-cdef-maps` keeps its id and asks what is still open.** With the tie
settled on the rule component, the question is how a framework the technology
provider did not map reaches the same rules: an SSP override, or a mapping
collection between frameworks, and which keeps rule-level precision.

**The copy-and-tailor convention is cited, and the questions stay open.** The
answers on the issue to `paper-comp-1`, `-3`, `-4` and `-5` describe one
convention: with no way to import a component definition, the SSP author copies
it, merges overlapping components, keeps the rules that serve the profile and
sets parameter values, with no link back to the source; rules reach the plan and
results through the SSP. Each question now states that convention and what it
leaves open.

**Question 4 is answered for component-first.** In the SSP every subject is a
component, the whole environment defaulting to This System, so the by-component
response names what the rule runs against. The cell moves from partial to
filled; its model and assemblies are unchanged.

**The runner convention is stated.** Question 5's cell and the shared
`proposal-3` question now say that each plan activity names its runner through
the proposed `assessment-asset-uuid` and its checks through `check-ids`, which
the committed assessment plan and results already show.

**The editorial standard is printed.** The questions page intro says a question
sits under the approach where it is most acute, not exclusively, and a question
every approach shares sits in the shared section.

**Not done, and why.** Questions 6a and 6b stay where they are: the issue says
findings were added to the assessment results example and an SSP example was
written, but neither file was attached, and the matrix is verified against the
committed examples. The "late, external" binding on question 2 stays unavailable
for component-first: it describes a separate mapping artifact reaching a rule,
which the schema does not allow, and an SSP override is the implementor's
response rather than that.

Large diffs are not rendered by default.

Original file line number Diff line number Diff line change
Expand Up @@ -1921,7 +1921,7 @@
document.documentElement.style.setProperty("--header-h", h.offsetHeight + "px");
}
/* --- the open questions: <div data-open-questions></div> ---------------
Four sections, one across all three approaches and one per approach, each
Four sections: shared questions and one per approach, each
a list of single-line questions that open into the reasoning behind them.

Questions and explanations are editorial analysis content in
Expand Down Expand Up @@ -1950,6 +1950,15 @@
det.appendChild(sum);
var body = el("div", "oq__why");
body.appendChild(el("p", null, q.why));
if (q.metadata) {
var fields = el("ul", "oq__fields");
q.metadata.forEach(function (name) {
var item = el("li");
item.appendChild(el("strong", null, name));
fields.appendChild(item);
});
body.appendChild(fields);
}
det.appendChild(body);
wrap.appendChild(det);
});
Expand Down
Loading
Loading