docs: add LibreSign brand and pronunciation guide

[vitormattos]

Summary

• adds a dedicated LibreSign brand and pronunciation guide;
• documents the official LibreSign spelling and pronunciation;
• explains “Libre” through practical software freedom and operational control;
• adds the current brand promise: Freedom and control for electronic signing.
• aligns the public visual-system wording with the completed canonical manual;
• links logo downloads directly to the canonical LibreSign/brand development release instead of serving copied website assets;
• links the current brand manual from the canonical brand release;
• keeps LibreCode institutional guidance out of the LibreSign page;
• identifies https://github.com/LibreSign/brand as the canonical version-controlled source;
• removes the Nextcloud share from the public brand flow;
• links the footer to the local LibreSign brand page.

Canonical brand architecture

The website is the human-facing publication layer. The canonical brand system — guidelines, SVG master, licensing metadata, reproducible manual source, governance, and history — is maintained in:

https://github.com/LibreSign/brand

The website should not become a second asset source. Public logo/manual downloads therefore point to generated artifacts from the canonical brand repository.

The latest brand release remains the mutable development/homologation channel. A stable v1.0.0 is intentionally deferred until external approval of the manual is complete.

Historical working material is not copied wholesale into the public repository. Only approved public guidance is published here.

Pronunciation playback and accessibility

The page now includes an accessible Hear pronunciation control.

The current implementation uses the browser's native Web Speech API as progressive enhancement, with no additional JavaScript library. It speaks the two parts separately so Libre can use a Romance-language voice and Sign can keep its English pronunciation. The button is only exposed when speech synthesis is available, is keyboard accessible, has descriptive context, and reports playback state through an aria-live region.

This is intentionally not treated as the final canonical audio: installed voices vary by browser and device. The preferred long-term reference is a short approved human recording, versioned in the canonical brand repository and published as a release asset. That follow-up is tracked in LibreSign/brand#27.

The written pronunciation remains available independently of audio so the feature does not create an audio-only dependency for accessibility.

The logo preview on the assets section now uses the canonical full-color logo intended for light backgrounds, fixing the previous low-contrast white-on-light application.

Test coverage and preview reliability

The pronunciation behavior is now implemented as a testable ES module rather than an untested page-specific script.

Coverage includes:

• language-prefix voice selection;
• Libre -> Sign playback order;
• busy/disabled control state during playback;
• successful completion state;
• error recovery;
• progressive enhancement when speech synthesis is unavailable.

Tests run with Node's built-in test runner through npm run test:js and a dedicated GitHub Actions workflow.

The preview workflow now uses wait-for-pages-deployment: true for both internal and fork previews. This keeps the shared site-preview deployment queue occupied until GitHub Pages has actually published the new commit, preventing a later PR preview from overtaking/cancelling the Pages deployment while the PR check already reports success.

The brand page labels also no longer force all-caps presentation; labels remain sentence case in the source and CSS.

[github-actions]

PR Preview Action v1.8.1-4-g2a6da11
🚀 View preview at https://libresign.github.io/site-preview/pr-preview/pr-497/
Built to branch main at 2026-09-25 20:00 UTC. Preview will be ready when the GitHub Pages deployment is complete.