[EI-6.1 — Experience Baseline & Accessibility Audit]

EI-6.1 — Experience Baseline & Accessibility Audit

Status: COMPLETE Date: 2026-08-08

Scope

Repository-level audit of the shared site experience, concentrating first on _layouts/default.html, _includes/meta.html, style.scss, and the navigation/terminal behaviors inherited by the corpus.

This is a code and structure audit. It does not claim to replace browser, screen-reader, or human usability testing.

Baseline strengths

Document language is explicitly declared (lang="en-US").
A viewport meta element exists.
Primary navigation has an accessible label.
The terminal input has an accessible label and is keyboard-operable.
Core content is rendered as HTML and does not depend on JavaScript to exist.
Images are globally constrained to the available width.
The main layout already includes mobile-specific rules.
The site uses visible text labels rather than icon-only primary navigation.

Findings

Blockers

Keyboard and assistive-technology users must traverse the masthead/navigation before reaching the main content on every page.

Action: add a visible-on-focus Skip to main content link targeting #main and style it so it cannot be obscured by the CRT overlay.

B2 — Motion effects do not honor reduced-motion preference

The global CRT scan/flicker layer and blinking terminal-style decoration animate continuously. The baud-rate display also changes automatically through JavaScript.

Action: add prefers-reduced-motion: reduce handling that disables decorative animation and smooth scrolling. Make the baud-rate animation settle immediately or remain static for reduced-motion users.

Important

I1 — Focus treatment relies too heavily on hover styling

Links have strong hover/active treatment, but the shared stylesheet does not establish an equally explicit :focus-visible system for links, navigation, and the terminal input.

Action: define a consistent high-visibility keyboard focus indicator that remains legible against dark panels and decorative overlays.

I2 — Viewport unnecessarily restricts user zoom behavior

maximum-scale=1.0 is present in the viewport declaration. Modern mobile accessibility should not impose an artificial maximum zoom scale.

Action: simplify viewport to width=device-width, initial-scale=1.

I3 — Main landmark uses redundant legacy ARIA

<div id="main" role="main"> works, but a semantic <main id="main"> is clearer and removes unnecessary ARIA.

Action: replace the wrapper with <main> while preserving the existing ID and classes.

I4 — Decorative CRT overlay sits at very high z-index

body:before uses z-index: 999. It has pointer-events: none, so it does not intercept clicks, but it can visually interfere with focus indicators and text clarity.

Action: ensure focus indicators and skip-link presentation render above the decorative layer, and disable the layer under reduced motion/high-sensitivity presentation where appropriate.

I5 — Terminal interaction is discoverable visually but has weak assistive feedback

Terminal output changes dynamically after commands, but the output region is not identified as a live/status region. A screen-reader user may not be informed when command output changes.

Action: use a restrained live-region strategy for terminal output or explicitly treat the terminal as decorative/optional. Avoid turning every line into noisy announcements.

I6 — Heading presentation forces uppercase site-wide

All h1–h6 are visually transformed to uppercase and carry relatively wide letter spacing. This supports the retro identity but can reduce scanning/readability, especially for long headings.

Action: retain the retro treatment selectively for structural/interface headings; reduce forced uppercase/letter spacing for long-form article headings and subordinate content where readability benefits.

R1 — Duplicate canonical link generation

default.html emits a canonical link directly and _includes/meta.html emits another canonical link. This is not primarily an accessibility defect, but it is unnecessary head duplication.

Action: keep one canonical implementation, preferably the absolute_url version in meta.html.

R2 — Google Fonts dependency should fail gracefully

The site loads JetBrains Mono from Google Fonts. The CSS font stack should remain fully usable if that request is blocked or slow.

Action: verify fallback stacks and avoid layout assumptions that depend on the webfont loading.

Seven primary navigation buttons are rendered in the masthead. Mobile rules allow wrapping, but the density can still dominate the first viewport.

Action: evaluate hierarchy in EI-6.4 before introducing a hamburger menu. Prefer a simple resilient HTML solution over JavaScript navigation unless testing demonstrates a need.

R4 — Decorative status bar creates repeated nonessential content

DIALING, NODE 80, SYSOP, and ANSI ON contribute strongly to identity but also add repeated chrome.

Action: retain if it remains visually subordinate; consider aria-hidden="true" for purely decorative status labels after confirming no meaningful information is lost.

EI-6 implementation backlog

EI-6.2

improve long-form heading readability
review article measure and spacing
distinguish interface typography from reading typography
verify font fallbacks

EI-6.3

add skip link
add :focus-visible system
remove maximum zoom restriction
convert main wrapper to semantic <main>
implement reduced-motion behavior
improve terminal assistive behavior
validate narrow-screen navigation

EI-6.4

evaluate seven-item primary navigation hierarchy
ensure History, The Loudmouth, subjects, archive, and portfolio remain discoverable without crowding primary navigation
preserve Writing rather than reintroducing Blog

EI-6.5

audit image alt text contextually rather than enforcing filename-based text
distinguish scans/artifacts from decorative imagery
retain privacy/media exclusions established in EI-1

EI-6.6

Automate only objective invariants. Strong candidates:

<html lang> remains present
viewport remains zoom-friendly
skip link and main target remain present
reduced-motion CSS remains present
primary navigation keeps an accessible label
known accessibility regressions are rejected

Do not attempt to automate subjective alt-text quality or visual hierarchy with simplistic string checks.

Decision

EI-6.1 is complete. No architectural blocker prevents implementation. The highest-value next work is EI-6.2 and EI-6.3: improve reading hierarchy while immediately correcting keyboard, zoom, semantic-main, and reduced-motion issues.