The Nature of Hypertext

By | 2014/01/14

Hypertext means more than just text with a bunch of links in it

Hypertext is something of a neglected subject these days. Everyone is talking about the Web, but nobody is talking about the class of thing the Web is: a hypertext environment. But to neglect this essential fact about the nature of the Web is to risk getting your approach to the Web seriously wrong.

Perhaps people associate the term hypertext purely with linking. But I would suggest that there is much more to hypertext than links. Hypertext is about the non-linear traversal of information spaces, and while links were the first tool created for this, there are now four key drivers of hypertext:

  • Search: Search engines allow readers to begin with their own ideas or questions and generate a collection of related texts. They also allow readers to generate a link from any piece of content whenever they want one. Thus, like links, the enable non-linear traversals of information spaces.
  • Social Curation: Social curation, in all its forms, from tagging to liking to tweeting, links creates a non-linear, non-authorial links across the Web.
  • Dynamic content: The dynamic generation of content, based on whatever information an individual supplies, creates content that is data-driven rather than linearly structured.
  • Links: Links remain the backbone of hypertext (the other drivers operate largely by generating links). But of the four, links are the one that is most in the hands of writers rather than readers, and therefore one of the writer’s sharpest tools in crafting hypertext.

Together, these four drivers account for the vast majority of the content and services that make up the modern Web. Unfortunately, many writers, particularly technical writers it seems, are ignoring all of them and continuing to create content as if for a linear static media such as a book, or for an isolated media such as a help system.

Content created in this way does not work well in a hypertext world. Nor do the tools and management structures designed to create such content work well for creating and managing hypertext. To understand why, let’s look at four key characteristics of hypertext:

Local over global

While the Web is a global phenomenon, it works because any given hypertext page is highly local. Each page tends to be short. It tends to treat a single local subject, often for a particular local audience. It links to topics on other subjects closely (locally) related to its own subject. The Web never offers you the 100,000 foot view. It does not show you the whole universe spread out before your feet (as some of the early encyclopedias attempted to do). Nor does it attempt to categorize and organize all of human knowledge. Its global scope is composed not of global structures, but of a web of untold billions of local pages connected by local links.

Wherever you are on the Web, you are always somewhere particular. You are never at 100,000 feet. You are always on the ground. Thanks to search and social curation you can explore your local area or you can leap half way around the world, but no matter how big a leap you make, you will always arrive somewhere local and specific.

This defies our old notions of how information is organized and found, but hypertext is a whole new way of organizing and finding information, and very clearly it works.

Item over collection

Hypertext prefers the individual item over the collection. This does not mean that there are no collections. The Web is full of collections, such as Wikipedia (a collection of encyclopedia articles), IMDB (a collection of movie information) and Amazon (a collection of goods for sale). But in none of these does the collection predominate over the individual item. Any article, movie entry, or product page on these sites can be added, edited, or deleted without anything having to be done to update the collection as a whole.

The collection is essentially the set of its pages, and it is the pages that place themselves in the collection, not the other way round. This blog post adds itself to the collection of Every Page is Page One posts. The same is true for forum discussion, Facebook posts, Tweets, LinkedIn discussions, or Google Plus posts. It is also true of the collection of technical help articles on sites like Stack Overflow that have been shown to provide better coverage of some popular APIs than their own documentation sets.

An important aspect of this stress on the item is that different people can add items to the collection at any time. There is never one climactic moment of publication towards which every contributor’s efforts must be directed and coordinated. Rather, each person contributes as appropriate by adding to the collection when there is information to add.

What this means, for purposes of tracking and managing, is that it is the pages that have revision histories and publication events, not the collection. There is no such thing as this year’s edition of Wikipedia (as there would be in the book world). There is only the members of the collection, appearing, changing, and disappearing as local conditions related to their own subject matter dictate.

Dynamic over static

Web pages are appearing, changing, and disappearing all the time. New links are being forged and old ones broken. The Web never stands still, and trying to create a stable citation to a resource on the Web is likely doomed to frustration.

This dynamic character is further enhanced by the many services that generate or aggregate content algorithmically. If you read an interesting story on Flipboard yesterday, don’t expect to be able to go back and find it today. It’s gone, and something else more topical is in its place.

Extraordinary measures are required to ensure that something will always be present and always at the same address on the Web. Actually  a great deal of content that is sticking around should probably have been taken down a long time ago. Permanence and immutability are not natural characteristics of hypertext, which tends much more to favor immediacy and mutability.

In the paper world, we actually had to go to great measures and considerable expense to compensate for paper’s lack of immediacy and mutability, as anyone who has had to deal with change pages can attest. But much of our management structure and practice is still based on the permanence and immutability of printed content, and is hard to adapt to the opposite properties in hypertext.

As readers we have the same expectation as everyone else: we expect what we read on the Web to be perpetually current. As writers, though, we have not adapted to the dynamic nature of hypertext. We still want our content to be fixed in time and immutable.

Bottom up over top down

Because it prefers the local over the global, the item over the collection, and the dynamic over the static, hypertext does not fit comfortably into top-down forms of organization.

Top down imposes a more global focus while inhibiting the full expression of a page’s local connections to other subjects.

Top-down emphasizes the collection rather than the item, allowing items into the collection only by blessing from on high, rather than by their own action. It thus inhibits individual contributions and funnels decision making through a single point, often introducing delay and compromising immediacy.

Top down favors the static over the dynamic, wanting each piece to be fixed in its assigned place in the hierarchy. It inhibits spontaneous addition, editing, and deletion, for fear of compromising the structure of the top-down edifice.

Finally, hypertext is bottom up because readers traversing a hypertext field are always at the bottom, always on a particular page with its particular local connections. They might parachute in from on high with a search or by following a curated link, or even an entry in a table of contents, but once on the individual page, all their subsequent navigation begins there, at the bottom. Every page is a hub from which further exploration proceeds bottom-up.

Adapting to a hypertext world

We live now in a hypertext world, but learning to create and to manage hypertext is no small undertaking. It should be clear from the above exploration of the nature of hypertext why putting your manual or help system up on the Web is no way to deliver Web-like hypertext content. It should also be clear that the change that is required goes beyond writing style or even tools. It involves a very different view of what the task of creating information is about and how it will be managed and governed.

Where should you start in the exploration of these changes? The first and most basic idea to grasp is this: Every Page is Page One. This is the implication of all the characteristics of hypertext, that it is local, individual, dynamic, and bottom up: Every Page is Page One.

If you wish to explore this further, I can recommend  a good book on the subject.


11 thoughts on “The Nature of Hypertext

  1. John

    Hi Mark

    Do you think that Apps pose a threat to hypertext in general? Perhaps companies would prefer to produce a self-contained App that will be limited but more fully under their control than allow a proper web/browser solution. I’ve seen corporate Apps that simply deliver PDFs, for example.

    1. Mark Baker Post author

      Thanks for the comment, John.

      Apps are certainly an attempt by publishers to subvert hypertext. The problem is, that an app does more to prevent people from finding its content than it does to keep readers from leaving it. Thus everyone who has an app also has to have a mobile website.

      I don’t know about you, but I don’t have any single-provider content apps on any of my devices. The only content apps I have are those that aggregate from multiple sites according to my preferences.

      Are apps winning the war against hypertext? I don’t think so. You might want to check out this article by Eddie Vassallo on the decline of tablet magazines:

      I have recently written an article for The Content Wrangler on this very subject. I’ll post a link when it appears.

  2. Chris Despopoulos

    Great post, and I think there’s lots of dollars and cents consideration to go into permanent vs temporal content and how it affects management, contribution, the pub worker’s value proposition, and whether or not content creation is becoming a commodity.

    I was with you all the way until you hit top vs bottom. I have two thoughts here. First, with hypertext there is no top or bottom per se — from any page you can navigate up or down. But also, for any conceptual framework there is a top and a bottom, and all this hypertext content is expressing concepts.

    So how do we reconcile this? The boundaries of a concept and its gravitational orientation are situational. They are established through social processes (conversation?) to suit a social context. We do this naturally all the time, and authoritative texts such as laws expend a lot of energy establishing boundaries and gravity. I would say one of the unique properties of hypertext is that it leaves text open to social adjustment of boundaries and gravity. But if it does, then you can’t say a “topic” is at the bottom any more than you can say it’s at the top. It all depends on the context.

    1. Mark Baker Post author

      Thanks for the comment, Chris.

      I think you raise a very astute point. I have argued back and forth with myself about the top-down vs. bottom-up distinction because mechanically, hyperlinks (except within something deliberately hierarchical like a TOC) go sideways. All pages are peers and all connections between them are therefore peer to peer. I could have chosen to say “Peer to peer, not master-slave”. But that is a bit more obscure, so I went with top-down vs. bottom up.

      But, as you point out, there is not just the mechanical relationship between pages involved here. There is also the relationship between concepts, in which one concept can indeed be thought of as being above or below another.

      One of the characteristics I describe for Every Page is Page One topics is that they stay on one level and use links to allow the reader to change levels as and when it suits them (from a workflow to an individual task, for example, or from a workflow to a big picture topic that puts the workflow in context). Such links are clearly going up and down in a conceptual sense.

      Overall, my decision to use top down vs. bottom up here was driven by the desire to draw as strong a contrast as possible between book/help organization (which is top down) and hypertext organization, to help readers appreciate how different hypertext is.

      As a standalone description of what hypertext looks like, though, one probably can’t come up with a better word than “web”.

  3. Pingback: The Nature of Hypertext | Technical Communicati...

  4. Pingback: Adapting to the modern web | Leading Technical Communication

  5. Pingback: Tech Writer This Week for January 16, 2014 | TechWhirl

  6. Manuel C. Reed

    While hyperlinking among webpages is an intrinsic feature of the web , some websites object to being linked by other websites; some have claimed that linking to them is not allowed without permission.

    1. Mark Baker Post author

      Thanks for the comment, Manuel.

      What you say is quite true, though it represents mostly failing businesses who do not understand how the Web works. For the most part, it seems to be newspapers, whose whole business model is really based on how far you can transport newsprint between 1am and when people get up for breakfast. It does not really work in a world in which you can transport the electronic bits of a news story around the world in the blink of an eye.

  7. Alex Knappe

    Hi Mark,
    while making some very insightful technical statements, your conclusion leads the wrong way.
    The nature of the web and hypertext is not constantly in the flux.
    The content web is highly dynamic and also highly static – and all shades of grey in between.
    Forums are one good example for that. While ongoing discussions are dynamic, older discussions may become static, as there had been found a solution, consense or weren’t of interest anymore.
    Some of those older discussions will be referenced by hyperlinks, if they’re containing content of interest.
    Some of them will be buried or archived.
    The same is true for other content on the web.
    Hypertext mainly gives you the possibility to create content that works top-down, bottom-up and on the same level, all at the same time.

    1. Mark Baker Post author

      Thanks for the comment, Alex.

      You are correct that there is a great deal of static content on the Web. In some ways you could see the Web growing like a tree: a ring of new growth each year around a static core of old wood that eventually succumbs to decay.

      But the dynamic nature of the Web does not really depend on new content appearing at old URLs. It depends on the mechanisms by which people navigate the Web. A page does not have to move or change for it to go from obscurity to viral hit and back to obscurity. The Million Dollar Home Page ( was a huge hit a few years ago. I don’t imagine it gets much traffic today.

      If you search the same subject today and again next month, you will likely get different results. Search from a different place, or logged in to Google as a different user, and you may get very different results.

      The thing about the Web is, it does not matter were anything is. It matters what things are about. That is what makes the Web dynamic: that things move in search space without changing location in real space. The Web is a dynamic filter, constantly moving things to the fore without every actually moving anything around.


Leave a Reply