Modeled connectivity
Follows explicitly recorded fibers and ports. Green does not prove a physical splice or live service.
FIBERATLASTry FiberAtlas Menu
UNDER THE HOOD
The checks.
The results. The limits.
We want you to see what backs this release. These are dated software checks with a defined scope. Real-world feedback is the next part of the work.
2,029automated tests passed
0failed or skipped
2signed Windows packages
FiberAtlas Beta 2.0 RC1 · 0.1.0-beta.20-rc1 · Source checks: September 11, 2026
Release verification
| Check | Result | What the evidence shows | Scope |
|---|---|---|---|
| Automated source regression | Passed | 2,029 tests passed, 0 failed, 0 skipped on September 11, 2026. The production build, local Beta Hub check, recovery check, and persistence benchmark also passed. | Source and test fixtures; not every Windows user journey. |
| Package identity & signing | Verified | Setup and Portable have valid timestamped signatures. The signer is John Philipe Kemmler. Both package hashes match the RC1 download manifest. | The exact two executables listed below. |
| Microsoft Defender scan | No threats reported | The signed Trial directory was scanned; artifact hashes stayed unchanged after the scan. | Point-in-time local scan, not a general reputation or future safety guarantee. |
| Windows installation & file launch | Passed | The signed Setup installed in Windows Sandbox. Opening an associated .fiberatlas file opened the requested synthetic project directly. | Focused Windows Sandbox scenario on RC1. |
| Uninstall & project preservation | Passed | Application, shortcuts, current-user file associations and running processes were removed. The existing project file remained byte-for-byte unchanged. | Focused current-user uninstall, not every upgrade or deployment scenario. |
| Trial delivery | Verified September 14 (UTC) | A confirmation email reached the test inbox, email confirmation unlocked the download, and a protected installer download started. | Email-to-download flow; not proof of an installer completing on another tester's device. |
HOW THE NETWORK LOGIC IS CHECKED
Known inputs. Expected outcomes.
Synthetic fixtures let us check failures as deliberately as successes. Expected assignments are specified independently of the trace result. These examples are fictional software tests.
| Scenario | Expected modeled result | What it checks |
|---|---|---|
| Three correctly connected services | 3 passed | Recorded paths reach their intended sources. |
| One explicit cut | 2 passed / 1 failed | A broken path must not look connected. |
| Two swapped assignments | 1 passed / 2 failed | Reachability alone is not the correct assignment. |
| One missing assignment | 2 passed / 1 incomplete | Missing intent stays incomplete. |
| 100 and 1,000 service fixtures | Each documented assignment checked | Bounded synthetic scale and cancellation cases; not a device performance promise. |
| Supported 1 × 32 PON | Explicit feeder-to-branch paths | Reverse tracing stays on the selected branch rather than lighting sibling services. |
Regression coverage also exercises recorded versus draft state, occupied and invalid connections, stale results, saved-run integrity, field-evidence import and history, persistence, and export disclosure. The 2,029 total is a software assertion count, not a count of field-tested networks.
Three results that stay separate
Modeled optical loss
Calculates with the selected wavelength, documented components, lengths and profile. Missing inputs remain unknown.
Measured evidence
Preserves user-entered or imported observations. It does not issue engineering or owner acceptance.
Where more testing is needed
- Broader Windows configurations, display scaling, and install/upgrade combinations.
- Shared drives, multi-workstation locking, and real concurrent users. RC1 support is bounded to local project storage.
- Physical printers and driver-specific layout behavior. PDF output has been exercised; a printer pass is separate.
- Long practitioner workflows with realistic projects. Synthetic scale tests are not a guarantee for all project sizes.
- Optical profiles, manufacturer data, owner criteria and real field evidence. The starter catalog is incomplete.
There is no overall field-validation, engineering-approval, or production-certification claim here. Tell us which environment or workflow you can help test →
Exact release identity
- Public label
- FiberAtlas Beta 2.0 RC1
- Embedded app version
- 0.1.0-beta.20-dev.1
- Source commit
6fe9a30d58b174a575badf062f689f99f58e4f1c- Build ID
fa-0.1.0-beta.20-dev.1-36975d4d9555a00a- Setup SHA-256
A89C7352E679A8D592880DC6F11F97A1B570A8D7B6A508670DB398288E148BB2- Portable SHA-256
8E14FEBF13BD96569047BFAF8D3300C411F252330F83B1CB789889411D31A1E3
The executable filenames retain beta.20-dev.1. Use the full hashes and build identity to match the published RC1, rather than the filename alone.
Read it. Try it. Challenge it.
Use the current user manual to try a small fictional project, then share what you find.
Prepared from retained internal source-check receipts, package manifests and Windows Sandbox observations, matched to the public RC1 package hashes. Prior-candidate results are not silently counted as fresh RC1 native tests.