Accessibility

Most accessibility statements describe an intention. This one describes what is measured, what it measures at, and what is known not to work yet.

The standard

WCAG 2.1 level AA. That is the standard referenced by Canadian accessibility legislation and the one this site is built and tested against.

What is checked on every build

None of this is asserted from memory. Each item is a check that runs before the site can be packaged, and a failure stops the package rather than producing a warning somebody may read.

Contrast, measured where the text actually sits. Every run of visible text is measured against the background genuinely behind it in the browser, walking up through transparent ancestors until an opaque one is found. Normal text must reach 4.5 to 1 and large text 3 to 1. This is stricter than checking a palette, because a palette pair can be correct while the pairing on the page is not. That check found a heading on the cookie banner rendering navy on navy at 1 to 1, present in the markup and invisible on every page.

Nothing scrolls sideways. Every page is loaded at eleven widths from 360 pixels to 3840 and checked for horizontal overflow, because a page that moves sideways on a telephone is the one fault a reader cannot work around. Tables and diagrams scroll inside their own frames so the page itself never does.

No required disclosure sits behind a click. Every disclosure block is checked for being collapsed, hidden, clipped by an ancestor, or positioned off canvas.

The reading column is measured in pixels at four widths, so it cannot quietly narrow below a comfortable measure as the layout changes.

What the site does

A skip link as the first focusable item on every page.

Keyboard operation throughout. Every menu opens on focus as well as on hover, the language switch and the carousel controls are real buttons, and a visible focus ring in gold appears on anything focusable. The dropdown parents are links with a separate control for the menu, so the parent page is reachable by keyboard and by finger rather than only by pointer.

Motion that can be stopped. The photograph strip advances on its own, so it carries a button with a word on it that changes between Pause and Play. It also pauses on hover and on focus, but a deliberate stop outranks the pointer, so moving the mouse away does not start it again. If your system asks for reduced motion, the strip does not move at all, and the hover effects keep their colour and shadow and lose their travel.

Touch targets of at least 44 pixels on every control. Where a control is a checkbox, the target is the label wrapping it, which is what a reader actually aims at. Writing this section is what caused that to be measured: the carousel buttons were 40 pixels, the card titles 25, the "Read" link 43.8 wide, and the consent label 43 tall. All four were four pixels or less short, which is the sort of miss that never shows in a screenshot.

Landmarks and headings. One H1 per page, headings in order, and the contents list on an article names every heading including subsections.

Text that resizes. Type is set from tokens and the layout scales in steps above 1700 pixels, so enlarging text or zooming does not break the columns.

A print stylesheet, so a page prints as a document rather than as a screenshot of a website.

No webfont. The typeface is Georgia, already on the device, so nothing depends on a font file loading and no text is invisible while one does.

No third-party request before consent. Nothing loads from another company until you choose, so nothing can inject an inaccessible widget into the page behind your back.

What is known not to work yet

Listed because finding out for yourself is worse.

Some photographs are still placeholders. Their descriptions are written and correct, but the pictures themselves are generated stand-ins on a small number of pages.

The French property is incomplete. Individual articles have not been translated yet, and until they are, some French pages describe an article in prose rather than linking to it.

Video and audio. There is none on the site. If any is added it will carry captions and a transcript before it is published, not afterwards.

Who to contact, and what they can actually change

Accessibility problems get fixed by the people who build the site, and on this site that is a small number of people rather than a department. A message about a barrier reaches the person who can change the markup, usually within the same week, and the change ships with the next build rather than waiting for a quarterly release.

What that means for you is worth being concrete about. A contrast problem, a missing label, a heading in the wrong order, a control that cannot be reached from the keyboard: all of those are changes to the source of this site and can be made quickly. A problem inside a document produced by an insurer, or inside a policy illustration, is not, because those are produced elsewhere and this site cannot rewrite them. Where that happens, say so and the information will be provided another way instead: read aloud on the telephone, sent as plain text, or set out in an email that your own software can handle.

There is no separate accessible version of this site and there will not be one. A parallel simplified site is always the version that stops being maintained, and a reader who needs it ends up with older information than everyone else. One site, built so it works, is the only arrangement that stays honest over time.

If something does not work

Telephone 514-875-9444 and say what you were trying to do and what happened. That is the fastest route and it reaches a person.

You can also write to info@ibcfinancial.com. Please describe the page and what stopped you rather than only naming the problem, because the same barrier can have several causes and the description usually identifies which.

A barrier report is treated as a defect and not as feedback. It is not necessary to be a client, to explain why you are using the site, or to name the assistive technology involved, though that last one is helpful if you know it.