Lend Me Your Compass, Cap!
Towards Informed Linking

As Mick Doherty--captain of Kairos  --has said on more than one ocassion, my job is one of the most exciting on our crew. Why? Because part of my job is to stand up on deck and just watch the ocean go by. I spend hours each week paying attention to the development of hypertext writing, design, and style on the Web... to think about how we do (or can) use this hypertextual environment for scholarship most effectively. One of the most thought-provoking issues that I've run into during the course of my Kairos  work so far is how our authors and others on the Web are measuring up to hypertext theory in terms of their linking styles and strategies.

Perhaps the most basic and exciting tenet of hypertext theory is that hypertext media and styles privilege readers in the writing/reading act; that is, the ability to choose (or bypass) links gives more creative and meaning-making power to the reader than do more linear forms of writing. How does writing on the Web currently measure up to this core of hypertext theory? Not well, I am sorry to say. Far too much of the linking we see in Web writing is what I like to call "blind linking." In this blind linking the author chooses a word or two that triggers for him a new node or train of thought. The reader then has the option of following or bypassing the link, but often with little idea of where that next node is going; the reader must either guess what the author is suggesting by linking from the word dog, or just follow the link and hope it pays off rhetorically and aesthetically, a follow-the-links method that can lead to another Web-reading phenomena, the "vlink syndrome."

In either case, the reader is forced to make rather uninformed decisions when navigating a hypertext, thus losing much of the meaning-making power attributed to the reader in hypertext theory. As Doug Brent mentioned in some e-mail concerning his recent submission to Kairos  (currently targeted for issue 2.1), relying on a few highlighted words for navigation information leads readers stumbling blindly around a hypertext -- hardly an empowering style. Readers bounce back and forth, hoping the links they follow pay off. I know I've spent many moments paused at my keyboard, wondering how I got from point A to point B based on that word I just clicked. Suddenly I am back to trying to figure out the author rather than creating meaning through my own reading decisions.

So, what can we do to better realize this reader-empowering goal of hypertext theory? How can we design our hypertexts to allow readers to make more informed navigating decisions? To give them the tools to chart their own ways through our works? I think possible solutions currently head in two directions: style and technology, two directions that are certainly not mutually exclusive.


As the Kairos  staff has always believed (despite some arguments to the contrary), there are many hypertextual styles on the Web, and we cannot dictate any One True Hypertext Style any more than we could with linear prose. I'd like to suggest, however, that we try to develop styles that provide as much information as is possible on each link. Perhaps that clarity can come through rewriting a sentence to include a more descriptive phrase to be highlighted, rather than linking from just a word or two. By pushing that </a> tag just a little further to the right, we can give far more useful descriptions of the link's destination. Just as writing teachers traditionally have asked for clear antecedents for pronouns like "this," we need to now look forward rather than back and ask for clear "linkacedents" that will let readers know where we are going.

A stylistic alternative to those expanded in-text links would be to cluster links at the end of each node in descriptive phrases. Instead of linking from the highlighted word "dog," an author might add the phrase "A discussion of how dogs are portrayed in popular movies" at the bottom of that node. And so on with other possible followups to the node at hand. The reader can digest a node as a whole, then decide where to go next, with less jumping between rapid-fire links and more flowing from one developed concept into the next. In addition, placing the larger content links at the bottom of each page would also free up in-text links for more referential purposes like definitions, clarifications, and examples. (See Brent's submission to Kairos  for his explanation and example of this stylistic technique.)

Again, as we continue to develop hypertext styles, the last thing we need to do is become dictitorial about the "best way" to write hypertext. These methods I suggested are just directions to explore; they certainly cannot fit the needs of every writer in every hypertext writing situation. Whatever the style, however, we do need to pay more attention to our readers, offering them informed options for navigating our webs.


Another option for providing this information is to redesign how browsers and HTML display link information. Currently, browsers show the link's destination URL. Karen McGrane Chauss' article in Kairos 1.2, however, demonstrates how the technology could improve the link information provided to readers. Placing a small Java script in each link tag, Karen uses an "OnMouseOver" option that lets her put descriptive text into that small window at the bottom of the browser screen, where the destination URL usually appears. Such a technique could be used to show what type of information lies at the other side of a link, allowing readers to make more informed decisions on which links to follow. Since this technique relies on a Java script that is readable by only the most recent versions of a few browsers, however, we cannot yet recommend its use to our other authors; in fact, a few colleagues have reported that the "OnMouseOver" tag will actually crash some versions and types of browsers.

Perhaps Netscape, Microsoft, and other browser companies will work on making this "OnMouseOver" information available in the next versions of their software. Just as we can now add descriptive "alt=" information to graphics tags, we could easily embed content information in hypertextual links. It is something to hope for, since one of the Web's biggest problems right now is blind linking. In addition to our own careful reflection on our developing hypertextual styles, such developments in technology would go a far way to bringing the reader-centered goals of hypertext theory to the World-Wide Web.

Again, I am not trying to dictate hypertext style or technology. Keep putting those <a href> tags wherever you'd like; freedom and diversity of style are wonderful aspects of the Web. I am strongly suggesting, however, that we all spend more time thinking about the rhetorical and epistemological power of those linking tags. How we link determines in part how our readers can interact with our texts. Let's remember, though, that if we want the empowered readers we talk about in our hypertext theory, we need to give them the tools to make informed navigational decisions; a compass is far more useful when the chart is reliable, too.

Greg Siering, Links Editor