Skip to content

VERIFICATION_AND_COMPATIBILITY.md

Latest commit

 

History

History
82 lines (55 loc) · 5.25 KB

VERIFICATION_AND_COMPATIBILITY.md

File metadata and controls

82 lines (55 loc) · 5.25 KB
Raw

Forest OS Verification and Compatibility

Authority: this document is the canonical record of what is verified, on what platform, at what scope, and what remains unknown. It absorbs and supersedes the verification claims scattered through the README history, 01_Docs/OPERATIONAL_STATUS.md, 01_Docs/PROJECT_LAUNCH_COMPLETE.md, and 01_Docs/FOREST_OS_PROJECT_HANDOFF_PACK.md §5 (see DOCUMENT_AUTHORITY.md).

Standing rule: every claim below carries a scope and a source. "Unknown" is a valid, preferred entry over an invented pass or failure. Source parsing is not runtime qualification.

1. Platform evidence

Platform Role Status Evidence
Ubuntu 26.04 LTS Development host (owner-reported) In use Owner report; no archived machine receipt in this repo
Ubuntu LTS / interim releases Support target Documented target Ubuntu publishes LTS (e.g. 26.04) and interim releases; Ubuntu is not a rolling distribution. 26.10 was a future release at documentation time — any parity claim for it is untested
Arch Linux / Omarchy Alternative target Documented target Dotfiles and environments remain separate from Ubuntu's; no parity receipt exists
Installable image Future ambition Not a deliverable No tested ISO/image exists; the repository is source only

Not claimed: host/guest parity, cross-distro parity, "upgrades cannot break the scientific stack", zero-footprint isolation. Containerization reduces dependency-collision risk; shared-kernel, driver, hardware, storage, and integration failure modes remain.

2. Component test inventory (source-level)

Static inventory of 05_Tests/ — 20 test functions:

Module Count Inspects
test_cli_wrappers.py 7 bash syntax, CLI help flags, exit codes
test_container_recipes.py 6 Containerfile directives, package manifests
test_desktop_launchers.py 7 FreeDesktop validation, SVG icons, server HTTP 200

Interpretation:

  • This is a source inventory, not a fresh execution result. A dated run receipt (passed/failed/skipped + host identity + commit) gets recorded below only when an actual run happens.
  • The desktop module can start the workbench web server on a fixed port and may skip if its validator is absent. Do not run blind test discovery on a workstation where the port is busy; a pre-existing service on that port is not your test result.
  • These tests do not qualify the five field-pipeline recipes end-to-end, nor the scientific correctness of canopy, biomass, tree-ring, or growth outputs.

Run receipts: (none recorded yet — unknown until a dated, host-scoped run exists)

3. Host tool verification (owner-reported, September 2026 scope)

From the handoff pack's verification table. Scope: one Ubuntu host, smoke-test grade (--help/CLI startup), not workflow qualification.

Tool Grade Receipt
BirdNET-Analyzer Smoke-tested (--help exit 0) Handoff pack §5 table, 2026-09-11
DeepForest Smoke-tested (v2.1.0 CLI operational) Handoff pack §5 table, 2026-09-11
pyDendron Smoke-tested (v1.7.5 CLI operational) Handoff pack §5 table, 2026-09-11
QGIS Installed (v3.42.3, APT) Handoff pack §5 table
GDAL Installed (v3.10.2, gdal-bin) Handoff pack §5 table

A --help check is not a complete ecological analysis test. Newer installed versions on the machine supersede any earlier planning values; do not downgrade receipts to match older notes.

4. Container recipes

Aspect Status Note
Base images :latest (both recipes) Unpinned; rebuilds may drift. Image digests not yet recorded — unknown
GitHub dependencies Unpinned revisions Recorded as a reproducibility gap (follow-up R3)
R package guard install_packages.R fails if required packages cannot load Useful guard; its printed checks are not archived build receipts
forest-sim FVS variants Unknown The compile helper's list has 20 entries and tolerates failures; the current Containerfile starts from an upstream FVS image and does not invoke the helper. The helper neither proves nor disproves the image's actual variant inventory — inspect the exact image and capture its executable list before claiming any count (20, 22, or otherwise)
Reproducibility posture Repeatable-in-practice, not reproducible-by-pin See README §7

5. Skipped checks and known unknowns

  • No independent synthetic scientific workflow has been run (follow-up R4).
  • No archived image digests or package-version manifests.
  • No machine inventory or hardware receipt is published (and none should be — private data).
  • Actual listening addresses of the workbench server and container services have not been verified from this repo's documentation task; see the README safety notes and follow-up R1.
  • The forest-tools planner's per-group verify commands exist in that separate project; their outputs have not been independently reviewed here.

6. How to add a receipt

When a check actually runs, append a row here with: date, host role (dev host / guest / CI), commit or image digest, exact passed/failed/skipped counts, and the receipt location. Do not edit historical receipts; supersede them.