Skip to content

Optimize Getting Started learning experience - #336

Merged
cdcavell merged 168 commits into
mainfrom
issue_work
Sep 14, 2026
Merged

cdcavell merged 168 commits into
mainfrom
issue_work

Conversation

@cdcavell

@cdcavell cdcavell commented Sep 14, 2026

Copy link
Copy Markdown
Contributor

Summary

Optimizes the docs/getting-started experience to make ASI Backbone Learning easier to navigate, scan, and evaluate without changing the underlying architectural guidance.

This update restructures the primary Getting Started pages around clearer entry points, shorter decision paths, and more consistent learning patterns while preserving the existing DocFX-compatible Markdown structure and destination links.

Changes

  • Optimized docs/getting-started/index.md

    • Added a clearer Start Here experience.
    • Condensed the architecture overview.
    • Made the five-part foundation easier to scan.
    • Simplified the Tutorial → Sample → Tests → Lab progression.
    • Reduced duplicate explanatory material.
    • Preserved the core governed-execution boundaries and existing navigation links.
  • Optimized docs/getting-started/adoption-personas-and-entry-points.md

    • Added a role/responsibility entry table.
    • Standardized persona sections around:
      • Typical question
      • Start with
      • Evaluate for
    • Added shared evaluation questions across roles.
    • Clarified that personas are navigation aids rather than strict job classifications.
  • Optimized docs/getting-started/find-your-path.md

    • Added a compact problem-oriented path selector.
    • Standardized each route around:
      • Start
      • Supporting material
      • Run and modify
      • Stop here when
      • Go deeper when
    • Reinforced the principle of using the smallest architecture that preserves the required boundaries.
  • Optimized docs/getting-started/learning-model.md

    • Added a concise learning-model overview table.
    • Clarified the progression from problem to tutorial, sample, tests, lab, comparison, and working implementation.
    • Simplified the distinction between canonical and alternative patterns.
    • Improved the explanation of framework independence and layered learning.
  • Optimized docs/getting-started/learning-path-map.md

    • Added a route-selection summary before the Mermaid diagram.
    • Simplified the explanation of recommended versus optional sequencing.
    • Converted the textual foundation into a scannable table.
    • Preserved the existing interactive Mermaid navigation and advanced-path links.

Design Goals

  • Improve first-time reader orientation.
  • Reduce cognitive load and repeated prose.
  • Preserve the existing problem-first learning philosophy.
  • Keep framework adoption optional.
  • Make tutorials, samples, tests, and labs easier to understand as complementary learning tools.
  • Maintain the distinction between recommended learning sequences and hard prerequisites.
  • Preserve current DocFX styling and navigation behavior.

Fix

  • Updated the Find Your Path link to reference the renamed Five-Part Foundation heading in getting-started/index.md.
  • Resolves the DocFX InvalidBookmark warning that was treated as an error by the documentation build.

Validation

  • All generated patches passed git apply --check.
  • git diff --check passed for each change.
  • Existing Markdown destination links were preserved.
  • Existing Mermaid click targets on the Learning Path Map were preserved.
  • No files outside docs/getting-started were modified.

cdcavell and others added 28 commits September 6, 2026 05:55
README.md tells readers to cite the Zenodo concept DOI, but CITATION.cff
carried no DOI at all, so GitHub's "Cite this repository" output and every
APA or BibTeX export produced from that file omitted the one identifier the
README asks people to use.

The CFF formatters behind that output read the top-level doi key only and
never fall back to identifiers, so the concept DOI goes there. It is also
the value that does not go stale: it resolves to the latest archived
version and matches both the README text and the badge. The identifiers
list then records the concept and version DOIs separately with
descriptions, keeping the distinction the README teaches visible in the
metadata itself.

The two entries under references are the AsiBackbone and
NetCoreApplicationTemplate repositories, which .zenodo.json already related
by DOI. They now carry those same DOIs, so the citation graph is the same
whether it is read from CITATION.cff or from the Zenodo deposit.

.zenodo.json declared its three DOI relations as scheme "url" with
doi.org URLs. Zenodo normalizes these on ingest, so the published record
already stores them as bare DOIs with scheme "doi"; this is a correction to
the source file rather than to the deposit, removing a discrepancy between
what the repository declares and what Zenodo actually holds.

ROADMAP's ongoing-maintenance item now names the version DOI alongside the
version and date, and records that the deposit mints it after the release.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
…eed link relative

`publish-docs.yml` runs `tools/validate-doc-metadata.cs` but did not list it
among the `tools/*.cs` path filters, so a change to that validator alone could
not republish the site. `docs-validation.yml` already lists it; add it here too.

The RSS autodiscovery link and the footer subscribe link hard-coded
`/Learning/feed.xml`, which 404s at the `localhost:8080` preview CONTRIBUTING.md
tells contributors to use. Render both with `{{_rel}}feed.xml` instead. Because
`_appFooter` is interpolated as raw HTML and is never re-parsed as a template,
the footer markup moves from `docfx.json` global metadata into the local
`_master.tmpl` override, and the template's header comment records the added
reason for the override.

`validate-doc-metadata.cs` now resolves the autodiscovery href against the
page's own canonical URL, so it still proves every page points at the published
feed while accepting the depth-correct relative form.

Fixes #267

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Adds docs/articles/2026/your-audit-log-is-not-evidence.md, a standalone
article arguing that a log line written after an operation narrates an
outcome rather than establishing what was decided.

The article works through four failures: the record captures the outcome
instead of the decision and its inputs, nothing binds the record to the
operation it describes, the record is written after the protected effect,
and integrity ends at whatever the log sink provides. It closes with
record-focused tests, explicit guidance on when an ordinary structured log
is the right answer, and a review checklist.

Registers the article in docs/articles/toc.yml and docs/articles/index.md
following the existing newest-first ordering and permanent year/slug URL
convention.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Enable and document the desired repository security posture for Learning,
including Dependabot security updates, secret scanning, push protection,
and an explicit main-branch ruleset.

Preserve required documentation, link, sample, and CodeQL checks while
requiring squash-only linear history, stale-review dismissal, resolved
review threads, and a pull-request-only emergency bypass.

Closes #287
@cdcavell
cdcavell merged commit f2c0921 into main Sep 14, 2026
11 checks passed
@cdcavell
cdcavell deleted the issue_work branch September 14, 2026 13:13
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant