The words Bush uses to describe his machine belie the key difference between hypertext on the web and hypertext as originally conceived: the memex was "for individual use" a sort of mechanized private file" a device in which an individual stores" an enlarged intimate supplement to memory, "Contrast this with most modern World Wide Web browsers, point-and-click programs that, not-withstanding a few forms-based interfaces, invited the user to surf the net without touching the content. Thus the net is a giant sprawling ecosystem in which the individual consumer is no more than a notch on each server's proverbial bedpost. Is point and click interacting, or, as one HotWired post suggests, just a higher form of channel surfing?

The weakness of modern hypermedia as an authoring environment has not gone unnoticed. In "The Missing Link: Why We're Doing Hypertext Wrong," Norman Meyrowitz argues that hypertext's potential as a revolutionary medium has been largely undermined by the lack of easy authoring opportunities in existing software (Meyrowitz 1989). Instead of the laborious tag-labeling of HTML, Meyrowitz asks for a system as simple as the "cut, copy, paste" paradigm used in almost all Macintosh and Windows applications. To create a hypertext link, Meyrowitz would allow a user to select a block of text or a graphic and choose "start link" from a menu, and then go to another point in another document and choose "complete link." The established links would last throughout the life of the document, and they would be bi-directional, allow a browser to go from document to document in either direction. Such authoring capabilities are expected to appear when Microsoft Windows '95 appears early next year, but even a program that allows users to write their own web pages misses an essential potential of cyberjournalism: if editors ceded some of their ability to control links to their readers, ideas embedded in hypertext could take on a life of their own. Imagine if a reader could build her own links, creating a "hotlist" not of favorite web spots, but rather a permanent thread through cyberspace identical to the pathways in Bush's memex. Then take the individual's thread a step further. Allow readers to create links from journalistic texts that connect to related spots on the net, commentary, other articles and myriad net resources. A cyber-reporter could not hope to know all the sites linked to the web that might interest their readers. Rather than rely on computer-driven, non-associative keyword searches, cybernewspapers should tap the vast human mind-power linked to the net to track down interesting links net-wide.

Needless to say, the wholesale proliferation of links to everything everywhere make hypertext almost irrelevant. If everyone creates links, one could create a document so heavily ÒlinkedÓ that every word was tied to its definition in five on-line dictionaries. Two innovations will be necessary to monitor the growth of on-line articles into "intelligent agents": link editors and link definitions. The editors will sort through suggested links and bestow the "stamp of approval" on those which actually lead to worthwhile information, while the link definitions will ease the smooth reading of a text.

In this scenario, readers could click on a word or a paragraph in a piece of cyberjournalism and create a new link to a site of their choice. On a regular basis, a link editor would read through the most-accessed documents and "edit" the links, establishing those which were most relevant as permanent parts of the story. The "unofficial" links would remain, coded differently (perhaps red instead of blue) so that the reader could distinguish between community input and "official" links.

The color-coding of these links points to the second essential element for cyberjournalism: link coding. Currently, links are often provided haphazardly in a text, sometimes for elaboration of a point, sometimes to the biography of a source and sometimes to the full document being discussed. In my first stab at hypercopy, I put a column I had written for The Chronicle on my home page. The piece was on the evils of highlighters, and included obscure references to Beavis and Butthead, Romeo and Juliet, Macbeth and the Communist Manifesto. All four cultural icons had linkable sites Beavis and Butthead had a fan page and newsgroup, while the three texts were available through FTP sites. I linked them all. Marx and Engels had nothing to do with my argument, other than an incidental example, but a reader surfing through my diatribe had no way of knowing whether the link was to the full text or to an elaboration of the role of highlighters in suppressing Communist revolution. Absurd though this example is, the problem is a fundamental one in cyberjournalism. Often, following a link can be slow, if not excruciating, and readers will quickly lose patience if they do not know what information they are getting and why. Successful hypertext will need to address this issue and develop a familiar set of symbols to make links more than novel accessories to reporting.

Notwithstanding the vagaries of what constitutes a worthwhile link, hypertext clearly has enormous potential for making stories more useful and informative. A politician's name might be linked to her platform from the last election, her campaign finance records and her voting record; a piece on school district problems might link to a page with a copy of the reassignment plan and a list of school board members. Similarly, coverage of President Clinton's State of the Union message might be paired with the transcript of his speech, along with a digital video for instant perusal. Recruiting this information, building links, and tracking down relevant information net-wide will require collaboration of the cyberjournalist and the link editor, versed in the net and able to both help the reporter construct an easily read web of links and track down useful links elsewhere.

But to take full advantage of hypertext, journalists will need to rethink what constitutes a story in cyberspace. Imagine different approaches to covering a residential fire in a small town. The story has dozens of angles: How did the fire start? Was there anyone in the house? Was anyone injured? Did the neighbors see anything? How quickly did the fire department get there? Were there any needless delays? Were there any teary-eyed children involved? Is there a good human angle? Depending on the circumstances, the reporter could write a lead about little Johnie's model plane that burned up in the fire or the new jet-power hoses that put out the flames in record time. Or, in hypertext, the reporter could let the reader choose. Herein lies the essential difference for cyberjournalists: rather than an inverted pyramid, most important to least important, or even a main narrative with hierarchical elaborations on certain points, true hypertext is a deck of cards shuffled and then pulled out according to the reader's own interests. Such a shift will require cyberjournalists to rethink the entire premise of reporting. In a print news story, the lead sets up the central tension of the story, providing the reader with essential information while setting up the questions an article intends to answer. In hypertext, there is no lead-just interlinking narratives in no particular order. So I might start by reading about Johnie's model plane, and then skip to the neighbor who called the police when she smelled smoke, then to a short biography of the firefighter who ran into the house to rescue Johnie's younger sister, and then to the details of the new hoses. Or maybe I'm a fire-fighting technology enthusiast, so I'll read first about how the new truck performed compared with the one the town replaced a few weeks earlier and contemplate the glitches in the computer-modulated water flow system. Instead of reading a summary of what the reporter believed were the essential elements, I can wander through the scene as a virtual witness. I choose the lead. There is no "angle."

Clearly, the typical small town, two-to-three beats and on-call reporter will not have the time to write a hypertext novel on every local fire, but such intimate detail is possible if interactive hypermedia is truly interactive. Imagine if the next door neighbor had grabbed a video camera and gotten footage of the fire truck as it arrived or if Johnie posted a message about his plane that burned up or if one of the fire-fighting technology enthusiasts posted a description of the new truck along with an explanation of potential problems. Everyone could be a reporterÑif the technology lets her.

For the issue of access is about more than just whether or not people will be able to get at all the information. Will they be able to put information out there on a equal footing? That is the "if" that Mitchell Kapor and Jerry Berman address in a New York Times op-ed piece:

None of the interactive services will be possible, however, if we have an eight-lane data superhighway rushing into every home and only a narrow footpath coming back out. Instead of settling for a multimedia version of the same entertainment that is increasingly dissatisfying on today's TV, we need a superhighway that encourages the production and distribution of a broader, more diverse range of programming.
(Kapor and Berman, 1993)

While Kapor and Berman's concerns may seem paranoid to the internet veterans who insist that no one will ever get away with charging for access to the net, the problem is very real if one begins to think about how access will beÑand has beenÑcoming into the American home. Prodigy, America Online and Compuserve all offer interactive possibilities such as newsgroups and email, but Prodigy has consistently censored newsgroup postings. Meanwhile, the advent of the "set top box" could put a very different kind of internet into the American living room. Armed with an infrared pointer to follow links and supplied with a melee of video-on-demand, how will the netsurfers of the future participate in lively on-line discussions, particularly if they don't even have a keyboard with their set top box? Indeed, if the net comes to the living room attached to a TV, interactivity is almost guaranteed to take a backseat. At a computer screen, the user is sitting up, attentive and typing. On the couch, he is much more likely to be eating popcorn, sipping beer and channel surfing. Certainly, being able to point to a score to get game highlights on demanded from SportsCenter is interactive but not in the dynamic way I would like to envision the future of cyberjournalism.

Many praise the ease-of-use of Web browsers without recognizing the inherent non-interactivity involved. Write Justin Hill in a HotWired post:

This is just another medium for telling the human story. It's exciting because it is intuitive and humanistic. Hypertext encompasses a wider variety of materials, and it allows you to explore, to move at your own pace, through content you select. I have even heard of plans to set up kiosks with touch-screens running Mosaic in public places. It would work because there isn't any extended literacy required. If you know how to read, you just point your way around a screen, following your whimsy.
(Justin Hill, Tue 25 Oct 94 9:09 PST)

True, users get to move at their own place through content, but as long as they only move without altering that content in any way, the net will be a one-to-many medium. Tellingly, Frank Daniels III, executive editor of the Raleigh News and Observer and grand pubah of Nando.Net, admits that the vast majority of Nando subscribers use the service for email, not for plugging through old stories or studying public issues in depth. Many-to-many, in the form of email, mailing lists, and newsgroups, is the true ÒinteractivityÓ of interactive multimedia. Dynamic links to tonight's TV are a very different story.

In follow-link only interactivity, as Laird Nelson points out, net surfers can't really follow their whimsy. Responsing to a post arguing that hypertext allowed the reader to explore a drama unfolding in a house as though walking through it, he writes on the alt.hypertext newsgroup:

Except of course we CAN'T move about as we wish. The only way for that statement to hold accurately would be if we could design our own links in the hypertext system as a way of emulating "moving about" in the house. The actual analogy between hypertext and the house in your scenario is more like a group tour: the docent can offer us choices ("Would you like to see the neo-Victorian bedroom or the Colonial kitchen?") but some areas will always be off limits.
Laird J. Nelson, alt.hypertext, 1 Dec 1994 15:50:07 -0500