Fahrner Image Replacement (abbreviated FIR) is a Web design technique that uses Cascading Style Sheets to replace text on a Web page with an image containing that text. It is intended to keep the page accessible to users of screen readers, text-only web browsers, or other browsers where support for images or style sheets is either disabled or nonexistent, while allowing the image to differ between styles. FIR is named for Todd Fahrner, one of the persons originally credited with the idea of image replacement.

Motivation

The typical method of inserting an image in an HTML document is via the tag. This method has its drawbacks with regards to accessibility and flexibility, however:

Browsers without support for images can't see the text. Adding alternative text using the alt attribute disallows HTML markup and causes problems with some search robots.

Using the tag to show text is presentational; many Web designers argue that presentational elements should be separated from HTML content by placing the former in a CSS style sheet.

Images referenced using an tag can't be easily changed via CSS, causing problems with alternate stylesheets.

Fahrner Image Replacement was devised to rectify these issues.

Implementations

The original FIR implementation described by Douglas Bowman used a heading, inside of which was a element containing the text of the heading:

Through style sheets, the heading was then given a background containing the desired image, and the hidden by setting its display CSS property to none:

It was soon discovered, however, that this method caused some screen readers to skip over the heading entirely, as they would not read any text that had a display property of none. The later Phark method, developed by Mike Rundle, instead used the text-indent property to push the text out of the image's area, addressing this issue:

The Phark method had its own problems, however; in visual browsers where CSS was on but images off, nothing would display.

Dave Shea's eponymous Shea method solves both of the issues earlier mentioned, at the cost of an extra :

By absolutely positioning an empty over the text element, the text is effectively hidden. If the image fails to load, the text behind it is still displayed. For this reason, images with transparency can't be used with the Shea method.