One of the greatest "sins" of web design, from an accessibility point of view, is attempting to control the font size of the display. Most of the people who do this seem to have extraordinary vision, because the hard-coded font sizes are often very small.

In large part, the use of fixed font sizes seems to derive from training that designers have received. Because the web is a relatively new method of disseminating information, many of the people currently designing websites were trained as print designers. In print design, exact colors, type faces, and type sizes are routinely controlled, so that the resulting document "looks" as much as possible as it was conceived by the designer.

Unfortunately, on the web, the designer has much less control over what the end-user sees. The colors being displayed, rather than being those desired by the designer, are those that are available on the site-visitor's monitor. They may be close to those specified, or may be quite different. The fonts displayed are those that the vistor has available, not those specified by the author/designer. The user may be visiting the site with an auditory browser (a screen reader or telephone-based browser) so that fonts and colors are irrelevant!

Even those who are not trained designers will often set their page layout by specifying internal features in terms of the number of pixels tall or wide an object is to be. Fonts, for example, can be specified in terms of points, pixels, or picas (12 points = 1 pica). Such specification may give the illusion of control to the author, but can make pages very difficult for the visitor who does not share the visual acuity of the author.

One approach to accommodating individuals with visual deficits is to provide a control on the page to adjust the font size. Typically, such a control will offer "small," "normal" (for the author), and "large" font on the page. Depending on the sizes selected for these sizes, this may actually meet the needs of many visitors. However, it is an imperfect solution at best.

The problem with author-provided font controls is two-fold. First, the visitor must locate the control on the page. There is no standard for where such a control should be, so the visitor with low vision must scan a page that is very difficult to see in search of a control that s/he hopes is available. Even if the control is in the same location on each page of a site, it is a new discovery for every site the user visits. This is an unreasonable burden on the individual with low vision. The second problem with author-provided controls is that they suggest very narrow consideration of the problems of visual sizing.

Consider the issue from the perspective of the individual with low vision. The need for large fonts for reading will not be limited to a single page, but will be consistent across the web and for all other print materials (including electronic print materials). Hence, it is likely that the visitor with low vision already has a general solution for making print more accessible. Using that known solution will be easier for the visitor, as well as allowing matching the display to the precise needs of the individual.

One means of making the web more accessible, for a person with low vision, is to install a "user defined" stylesheet (this stylesheet uses the standard CSS definitions to modify pages to the user's needs) on the indvidual's browser, so that all pages that allow such stylesheets will be displayed appropriately.

Unfortunately, if the page uses hard-coded font sizes, such a stylesheet cannot modify the display, and the page becomes inaccessible.

What to do, what to do?

While the web specification and standards allow web authors to design in ways that are not accessible, they also provide alternatives that make pages more accessible. In this case, the solution is to assure that all font specifications are written using "proportional" units. Where pixels are the fixed dots on the screen, and points are (theoretically) 1/72 of an inch, other sizing methods are also provided. These include percentages, which specify that the item is to be displayed in some percentage of the container standard. In font sizes, this is a percentage of the base font size of the page. If the size of the base-font changes, the other fonts change proportionally. Other portions of the page can also be specified in percentages. Floating boxes like the "Quote for Today" on this sites home page can be specified as a fixed percentage of the page (or container) width, so that when the page is resized, the text box changes in a predictable manner. Ems and exs are also relative font size specifications. Where the em is based on the distance between the top and bottom of an uppercase letter, the ex is based on the height of the lower-case "x."

Because of the details of font design, all 12 point fonts will have similar em heights (in theory 12/72 of an inch), but the ex-height can change between different type faces.

The key point, though, is that when font sizes are specified in percentages, ems, or exs, the user can increase or decrease font sizes as needed, either using a custom style sheet or the keyboard commands available in most browsers.

For the designer, the message should be that it is the needs of the reader, not the aesthetic of the designer that should control design. Feel free to specify size relationships (this should be twice as big as that), but, unless you want to make your design difficult to use for anyone with limited visual acuity, don't try to specify just how big the fonts should be.