Publishers of technology books, eBooks, and videos for creative people

Home > Articles > Web Design & Development > Usability

  • Print
  • + Share This
  • 💬 Discuss

This chapter is from the book

Basic Approach (Overview)

The i3Forum layout is designed to deliver a crisp, punchy brand identity with a minimum of fuss, and its XHTML is equally straightforward. It is composed of two XHTML tables, both centered, and both enhanced and controlled via CSS. The first table delivers the navigational menu; the second provides the content [Figure 8.2].

Figure 8.2Figure 8.2 The template we'll build in this chapter, with CSS turned off and borders turned on. Note the slightly thicker line between menu and content areas, where two separate tables meet.

The XHTML for the tables will be shown in the pages that follow. But a preliminary question might already have occurred to you. Traditionally, such layouts would use a single table, with rowspans and colspans juggling the various rows and columns. If we used Adobe ImageReady to automatically slice and dice the Photoshop comp used to design the site (and to sell the design to the client), ImageReady would render the entire page in a single table. So why have we used two tables?

Separate Tables: CSS and Accessibility Advantages

If you skipped Chapter 7's ("Tighter, Firmer Pages Guaranteed: Structure and Meta-Structure in Strict and Hybrid Markup") discussion of "div, id, and Other Assistants," you might want to glance at it before going any further. Breaking our layout into two tables allows us to harness the power of the id attribute to do the following:

  • Streamline the CSS we'll create in Chapter 10

  • Provide certain accessibility enhancements

  • Structurally label each table according to the job it does, making it easier to some day revisit the layout and replace presentational XHTML tables with divs styled via CSS

The Table Summary Element

In addition, breaking the layout into two tables allows us to add a summary attribute to each:

<table id="nav" summary="Navigation elements" ... etc. >
<table id="content" summary="Main content." ... etc. >

The summary attribute is invisible to ordinary desktop browsers like IE and Netscape. But the screen-reading software used by non-sighted visitors understands the summary attribute and will read its value aloud. In our case, the screen reader will say "Navigation elements" and "Main content." Well-designed screen readers allow users to skip the table if they don't think it will interest them. Writing table summaries thus forms a good accessibility backup strategy to accommodate users who might miss the Skip Navigation link described two paragraphs from now.

Page Structure and id

We've assigned an id attribute value to each table according to the structural job it does—navigation or content. Doing so now allows us to later write compact CSS rules that apply to an entire table, avoiding classitis and divitis (defined and discussed in Chapter 7).

It also allows us to provide a Skip Navigation link in the top of our markup.

The What and Why of Skip Navigation

As its name implies, the Skip Navigation link allows visitors to bypass navigation and jump directly to the content table by means of an anchor link. The id attribute whose value is "content" provides the anchor to which we link:

<div class="hide"><a href="#content" title="Skip navigation." accesskey="2">Skip navigation</a>.</div>

Skipping navigation is not an urgent requirement for most sighted web users, who can focus their attention on particular parts of a web page simply by glancing at those parts and ignoring other parts that don't interest them.

But nonsighted visitors who are using screen readers experience the web in a linear fashion, one link at a time. It frustrates such users to endure a constant audio stream of menu links each time they load a page of your site. Skip Navigation lets these users avoid this problem.

Skip Navigation can also help sighted readers using non-CSS-capable PDA browsers and web phones avoid tediously scrolling through a fistful of links every time they load a new page. Finally, Skip Navigation can benefit sighted users who are physically impaired, although the method is not perfect. (See the later section titled "accesskey: Good News, Bad News.")

Skip Navigation and accesskey

Our Skip Navigation link enables visitors who are using nonvisual or non-CSS browsers to jump directly to content in the second table, whose id attribute name (and thus whose anchor link) is "content":

<table id="content" ...> etc. 

In these nonvisual or non-CSS environments, the link is readily available at the top of the page [Figures 8.3 and 8.4]. You'll create a rule to hide the Skip Navigation link in CSS-capable browsers in Chapter 10. (If you're the impatient type, we've also included it here.)

.hide  {
  display: none;
  }

Figure 8.3Figure 8.3 In a non-CSS browser (or a CSS-capable browser with CSS turned off), the Skip Navigation link is clearly visible at the top of the page.

Figure 8.4Figure 8.4 The visible Skip Navigation link in context—in our layout with CSS turned off.

Because of this CSS rule, visitors who are using modern browsers with CSS turned on will not see the Skip Navigation link—but most of them do not need to see it because they do not require Skip Navigation functionality for the reasons discussed in "The What and Why of Skip Navigation." Screen readers that ignore CSS will merrily read the content of the div, thus informing nonsighted visitors that they can avoid the boring recitation of the other links. (Alas, some screen readers obey CSS even though their users can't see it.)

There's an exception to every assumption, of course. A person who has impaired mobility, viewing the site via a CSS-capable browser, might desire to skip the navigation area and jump directly to content. Most web users who have impaired mobility can see an entire web page at once (the exception being users who are visually and physically impaired). But to navigate that page, impaired users employ the keyboard or an alternative, assistive input device. Tabbing their way past unwanted navigation links could be a nuisance, or worse.

How can we help these users skip navigation if they can't see the Skip Navigation link in their browser? We've provided that option via the accesskey attribute, which works even when the Skip Navigation link is invisible in the browser. Alas, the method is imperfect, as discussed next.

accesskey: Good News, Bad News

The accesskey attribute to HTML/XHTML enables people to navigate websites via the keyboard instead of a mouse. To assign an accesskey to an element, you simply declare it, as in the earlier XHTML excerpt, which we reprint here with the relevant attribute and value highlighted in bold:

<a href="#content" title="Skip navigation." accesskey="2">Skip navigation</a>.

In our markup, we've assigned the Skip Navigation link an accesskey of 2. Therefore, to skip navigation, the visitor simply presses 2 on her keyboard. As is often true of accessibility enhancements, the required markup is easy to write and has no effect on the site's visual design. In this case, that's both good and bad.

For how does the visitor know to press 2 on her keyboard? No widely used browser displays accesskey letter assignments. Neither do most little-used browsers.

accesskey and iCab

As of this writing, only iCab [Figure 8.5], a Macintosh browser, visually displays accesskey letter assignments. Most web users are not Macintosh users, and most Macintosh users are not iCab users. Making matters worse, iCab cannot show the accesskey assignment when the Skip Navigation link is hidden via CSS. As this book goes to press, iCab still does not support much of CSS1, the W3C's first CSS recommendation, published way back in 1996. In short, although iCab is an interesting browser and its commitment to supporting HTML 4 is impressive (and no, we're not being facetious: iCab's HTML 4 support is superb), iCab is not going to solve the world's accesskey problem.

Figure 8.5Figure 8.5 Of all the world's browsers, only iCab for Macintosh (http://www.icab.de/) displays our accesskey of 2, cuing the user that she can skip navigation by pressing 2 on her keyboard.

Two Utopian Possibilities for accesskey

Clearly, the majority of users who might benefit from accesskey have no way of knowing which accesskey letters or numbers to press; therefore, they cannot benefit from it. Because of that, including accesskey in your markup is somewhat idealistic.

If the W3C would recommend standard accesskey assignments for universal functions like "skip navigation" (and if designers and developers would follow those recommendations), users would always know which keys to press. That would be a good thing.

Alternatively, browser makers might decide to beef up their accesskey support by visually displaying accesskey values if the user decides to turn on this accessibility option in his preferences. IE for Windows provides an accessibility option allowing users to ignore font sizes on any web page. It might also add an option to Always Show Accesskey Values.

We must admit it feels rather Utopian to hope that the W3C will standardize accesskey shortcuts any time soon, and it also feels Utopian to hope that any major browser vendor (let alone all of them) will devote engineering time and resources to an always-visible-accesskey option. Nevertheless, we continue to use accesskey. Some users might view source to see which accesskey values are in use on a page and thereafter use the appropriate keys to navigate. We hope things become easier for these users soon.

Additional id Attributes

In addition to the primary id attribute names (nav and content), in our first pass at the site's markup we also assign unique id attribute names to each cell of the navigation table. Two cells should suffice to make the method clear:

<td width="100" height="25" id="events"><a href= "events.html">Events</a></td>
<td width="100" height="25" id="schedule"><a href= "schedule.html">Schedule</a></td>

We also assign unique id attribute values to each of the two primary divisions of the content table, namely the sidebar (id="sidebar") and primary content (id="primarycontent") areas. Next, with much data removed for clarity, is the shell of the content table; id attributes and values have been highlighted in bold:

<table id="content" etc.>
<tr>
<td width="200" id="sidebar">
Sidebar content goes here.
</td>
<td width="400" id="primarycontent">
Primary content goes here.
</td>

For good measure, we slap an id attribute name on the secondary rows of the navigation bar. Thus, the second row of navigation "buttons" has the following id value:

<tr id="nav2">

And, as you might expect, the third row of navigation "buttons" has this id value:

<tr id="nav3">

How Much Is Too Much?

The latter two id attribute names (nav2 and nav3) aren't required for this layout's purposes, but they might come in handy one day, in the event a redesign is required. Should we include them or not? Including them now adds a few bytes to our XHTML, and we might with equal merit have chosen not to do so.

If, on the final site, the navigation bar lives in a separate Server-Side Include file (or in a unique record managed by PHP, JSP, ColdFusion, or ASP), the client could easily edit that file at any point in the future, changing the entire site by adjusting a single file. If the client plans to use server-side technologies, it might be silly to include nav2 and nav3. On the other hand, if no server-side technologies are used and the menu markup is manually repeated on every page, it might be safer to go ahead and include nav2 and nav3 to avoid potential search-and-replace errors in a future redesign. And that is what we've done.

  • + Share This
  • 🔖 Save To Your Account
Designing With Web Standards

This chapter is from the book

Designing With Web Standards

Discussions

comments powered by Disqus