1.4.4 Resize text (AA)

The intent of this Success Criterion is to ensure that visually rendered text, including text-based controls, can be scaled successfully so that it can be read directly by people with mild visual disabilities, without requiring the use of assistive technology such as a screen magnifier. Users may benefit from scaling all content on the Web page, but text is most critical.

When content is encoded to have specific, fixed size (either for text, images or in some cases layout features such as column widths or padding), users may struggle to perceive the content, due to either sensory difficulties or device limitations (e.g. vision impairments or having a small screen).

Depending on their specific needs, users may use their browser's built-in full-page zoom functionality (which zooms all content in equal proportion). However, if this has not been accounted for during the design and development of the content, it can result in problems such as elements being cut off and overlapping content/controls (which can make a site completely unreadable/unusable).

While in the past browsers only allowed users to resize text, by default modern desktop browsers now use what is called "full-page zoom". In this mode, all content (text, images, overall page layout) is zoomed, regardless of how content dimensions were defined (both absolute measurements, generally using px, and relative units, such as em, are scaled the same way).

When content is designed using specific fixed dimensions, it will not adapt to the user's browser window or to any changes in full-page zoom levels. If the width of the content exceeds the width of the window, users will have to scroll both horizontally and vertically. This introduces significant scrolling burden on users, who now have to pan across the site in order to be able to perceive it, and risks them missing important content in the process.

Certain desktop browsers (Firefox, Safari on OS X) offer users the ability of only zooming/scaling text (by effectively changing the base font size of 1em), leaving all other content dimensions unchanged (unless these were defined by using a measurement unit relative to the base font size, e.g. container dimensions defined using em as well).

Further, users who are already using screen magnification software are already zooming and panning across the screen, so this scrolling adds an extra layer of zooming and panning, causing considerable loss of time in navigating the site, increased cognitive load, confusion and making it likely that important content will be missed. In addition to full-page zoom, Internet Explorer provides users with a dedicated "Text Size" menu option (with 5 size options: "Largest", "Larger", "Medium", "Smaller", "Smallest"). However, this text size setting only affects fonts defined using relative units (such as em) - text sized using absolute measurements (e.g. px) remains unchanged.

Note: this success criterion does not explicitly specify what method of text resizing is supported. If a site works perfectly fine when using full-page zoom (i.e. no elements overlap/are cut off), then it is a pass - regardless of whether or not zooming results in users having to scroll (though there may be exceptions if zooming results in a large number of scroll bars within the page, making the content hard/impossible to use).

Note: at a high level, testing this success criterion involves zooming to 200% on a browser window sized to a mid-range resolution (currently 1024 x 768), as well as testing in a browser that still offers text-only resizing options (see below) and changing the text size to 200%, and ascertaining if as a result of zooming/resizing any content overlaps, is cut off, or is otherwise rendered unreadable/unusable. If zooming/resizing results in horizontal scrolling, testers will need to use their judgment to decide if this horizontal scrolling is excessive and likely to cause problems to users (for instance, if text resizing results in various containers within a page to start featuring horizontal scrollbars, making the overall page difficult to read/use due to the number of scrollable areas that are now present). As such, there is an element of subjectivity with this how this success criterion is tested/evaluated.

Note: if a page contains no text (which is an uncommon/unlikely occurrence), this criterion is marked as not applicable.