Accessibility

The standard being aimed at

The target is the Web Content Accessibility Guidelines, version 2.2, at Level AA. Two decisions go beyond it deliberately: buttons and button-shaped controls are at least forty-four pixels in their smallest dimension, which is more than the Level AA target size requires, and the focus indicator is never suppressed anywhere. Links inside prose and dense reference lists keep their normal text size, with the spacing between them held at or above the Level AA minimum instead.

This is not a conformance claim. Full conformance to a standard is a formal statement backed by a formal evaluation of every page against every success criterion. That has not been done here. What follows is a description of testing, with its gaps named.

What has been tested

Keyboard navigation

Every page was walked from the first tab stop to the last using the keyboard alone. Every link, button, form control and disclosure can be reached and operated without a pointing device, the focus ring is never removed, and the tab order follows the reading order. A skip link is the first focusable element on every page and moves focus to the main content.

Focus management in dialogs

The search panel and the narrow-screen navigation both use the native dialog element, so focus containment is handled by the browser rather than by hand-written code. Escape closes both. In the search panel that takes one line of script, because a search field consumes the first Escape to clear itself, which would otherwise destroy the query and leave the panel open. Both return focus to the control that opened them when they close, which browsers do not guarantee on their own. Both were checked by opening and closing them with the keyboard alone.

Heading order and page structure

Each page has exactly one first-level heading and its heading levels descend without skipping. Headings were checked page type by page type, and heading level is chosen by position in the document outline rather than by how large the text should look.

Landmarks and labelled regions

Every page exposes a banner, a main region, a content-info footer, and named navigation regions. Where more than one navigation exists on a page, each carries its own accessible name, so a screen reader user can tell the breadcrumb from the chapter list from the on-this-page list.

Reflow at 320 CSS pixels

Pages were checked at 320 CSS pixels wide, which is the reflow requirement in the standard, and at 400 per cent zoom. Nothing requires horizontal scrolling of the page itself. Wide tables scroll horizontally inside their own container instead of widening the page, and long Greek transliterations, checksums and web addresses are allowed to break rather than push the layout out.

Colour independence and contrast

Nothing on this site conveys meaning by colour alone. Every evidence role, review status, comparison and change type carries its own words, and the badges add a text glyph as well. Body text, headings and interface text were checked against the contrast requirement for their size, and pages were reviewed in greyscale to confirm that nothing became ambiguous.

Reduced motion

There is very little motion here: no carousels, no parallax, no animated illustrations, no auto-playing video, and nothing that appears or moves as you scroll. What motion exists is confined to things you operate, such as a panel opening or a button responding to a press, and to scrolling smoothly when you follow a link to a heading further down the same page. When the operating system reports a preference for reduced motion, every one of those movements is removed, replaced by a plain fade or by nothing at all. The colour changes that show you where the pointer is are kept, because taking those away would make the interface harder to follow rather than gentler. Both versions are checked automatically on every build, and the still version was also reviewed with the preference switched on.

Greek and Hebrew language attributes

Original-language text is marked with its own language attribute, so a screen reader announces Greek with Greek pronunciation rules and Hebrew with Hebrew ones rather than reading them as broken English. Hebrew is set right to left and isolated so that it cannot corrupt the direction of the surrounding sentence. Every occurrence carries a transliteration and a gloss in English, so no argument depends on reading either script.

Print output

Printed pages were checked on paper. Navigation, dialogs, search controls and the video embed are removed. The argument, the Scripture quotations, the citations and the source list remain, disclosures are opened so that nothing is lost inside a collapsed section, and external link destinations are printed after the link text so a printed page remains checkable.

Automated checks

Automated checks with the axe rule set cover representative pages of each type: the homepage, a case section, a passage page, an index page, the corrections form and this page. They are part of the project’s test suite rather than a one-off audit, so a regression shows up in a test run rather than waiting to be noticed by a reader.

Automated checks and manual review are not alternatives to one another. The automated checks run on every build and catch regressions quickly; the manual work is what establishes whether a page is usable. Both were done, and neither on its own would be enough.

Known limitations

No independent audit

This site has not been audited by an independent accessibility specialist, and no formal conformance evaluation has been carried out. Everything on this page is the result of testing done by the people who built it.

Limited screen reader coverage

Screen reader testing has been limited in scope. Behaviour with combinations that were not tested may differ, particularly around the tables, the disclosure sections and the search panel. Reports about a specific screen reader and browser combination are especially useful.

Wide tables scroll sideways

Several tables carry comparisons that genuinely need three or four columns. On a narrow screen those tables scroll horizontally inside their own container. That is a real cost for some readers, and the compromise was chosen over splitting a comparison into pieces that can no longer be compared.

The video player is a third party

The overview video is hosted on YouTube and its player is not under this site’s control, so its keyboard behaviour and its own controls cannot be corrected here. Nothing on this site requires watching it: the full transcript, the chapter list and the written case cover the same material.

Original-language rendering depends on your system

Greek and Hebrew are rendered with the fonts your device provides for those scripts. On a system with poor coverage the text may render with substituted glyphs. The transliteration and the English gloss beside every occurrence exist partly for this reason.

Automated checks find only part of the problem

Automated tools reliably catch missing names, contrast failures and invalid markup. They cannot tell whether a link makes sense out of context, whether a reading order is meaningful, whether an error message helps, or whether a table caption describes the table. That is why the list above is mostly manual work, and why it is described as testing rather than as conformance.

Without JavaScript

The whole case is readable with scripting turned off or unavailable. Pages are rendered as HTML, navigation is plain links, and the reading experience does not depend on client-side code.

Four features degrade rather than disappear. The search panel becomes an ordinary link to the search page. The narrow-screen menu falls back to the footer, which carries the same structure. The corrections form is a real form that posts to the server, so a correction can still be sent. The video becomes a link that opens it on YouTube in a new tab, rather than a player that loads in place.

Four controls are absent rather than degraded: the filters above the source library and the Scripture index, and the reading-progress panel and the print button on the case map. Each one only narrows, records or reformats something the page already shows in full, so every list they act on is rendered complete and unfiltered and nothing is hidden by their absence. What is missing is the means to narrow a list, the note of which parts you have opened, and a button that does what your browser’s own print command does.

Reporting a problem

Accessibility reports are treated as corrections, which means they go into the same queue as a factual error and are published in the changelog when they are acted on. Several pages on this site exist in their present form because the original material was not accessible, and those changes are recorded as accessibility revisions.

Please say what you were trying to do, what happened, and which browser, operating system and assistive technology you were using, including version numbers if you have them. A page address helps a great deal.

Report an accessibility problem

That link opens the corrections form with the accessibility type already selected. No email address or account is needed, and nothing about your report is published unless you choose it. This page was last reviewed July 29, 2026.