Coleridge – An Open Source Xanadu

This is a visionary text written in 2001 before Web 2.0, Youtube, Facebook and blogging. It predates much of the current debate around the Semantic Web (Web 3.0) and is in line with the ‘fuzzy’ as opposed to ‘neat’ view of the emerging Web’s structure.  I circulated it amongst friends and even sent a copy to the Shuttleworth Foundation. My comments looking back from 2010 are in red. DRL

Coleridge, an Open Source Xanadu

[Published under a Creative Commons By NC  license, please see About page.]

Map showing location of Xanadu

Clearly, new forms of visual representation and spatial organisation of the web (the ordering of it’s growing content) will be the next, big wave, and as the net inflates, the visual mapping of information and the creation of new forms of data landscape or hypermedia will become increasingly important. Imagine a giant library filled with books of all description.

If it weren’t for John Dewey who invented the Dewey system of classification, one would never find the same book twice.

The mechanical nature of the web has meant that information is seemingly ordered and open to enquiry, however, precisely because there is no standard means of ordering information, no one particular catalogue of the web, this appearance is misleading. Intelligent agents and context sensitive searches / search engines are one way of traversing. Labelling. Meta-tags are another, but what we all really need is a better visual interface, some form of graphical organisation that overcomes the difficulty faced by a system in which subject, title, author and content do not necessarily correspond in convenient and accessible categories.

When Ted Nelson envisaged the Xanadu Project in 1960, he foresaw the possibility that the links between things (objects and information) would become more important than the things themselves. In fact, we would all end up publishing our links, some of which would actually be sold. published or licensed, enabling a form of income to accrue from the royalties gained by the sale of hypermedia.

While this was a bad idea in terms of its commercial appeal, and the  Xanadu Project was (and still is) the biggest vapourware project in history, its success in marketing the concept of hypertext places it near Internet ground zero, at very least, it is the basis for the thinking surrounding hypermedia and the world wide web.

The question then, is Xanadu relevent today? Can we make the links and connections to things seem important again? [This question refers to the as yet unnamed, Semantic Web]. Is there a missing Third Dimension to the Web of Information?

Three types of spatial organisation

The answer to this question came as a result of a short journey of discovery (one Saturday afternoon in 2001), in which I traversed the same problems encountered by both, Dewey and Nelson. The result was astonishing, because it became clear that we were doing something patently wrong by web browsing, and that there was an essential ingredient missing between web-site, and web-browser, in other words, a system of mapping and visualising links which does not treat the web as some form of encyclopaedia or a giant book, but rather as a relatively open space, perhaps a stellerium in an open library, prone to dusty nooks and crannies in which one may witness both the decay of the old and the marvel of the new.

In a library, it is easy to see which books are older, and frequently read, and which books are newer or unread.

There is a form of virtual decay on the web, but this is merely an illusion created by the interface which presumes that links have either been used or not, a particular direction is either followed and visited or remains unvisited.

We make bookmarks, and there is vicarious history, but  nothing showing you how one site is related to another.

For instance, is the information older or newer?

Was this book published today or ten years ago?

How many pages and versions have been published?

Am I nearer to the source than I was at the beginning, or am I getting colder?

Signpost in Antarctica

The subjective sense of adventure within the walls of a library is the result of various cues which help one on in ones intellectual journey. An old dusty book found in a basement or attic will be different from a polished journal found near a stair well; likewise, one expects to find different levels and forms of media, and our expectations vary according to the perception of commodity value, typography and so on.

For instance, compare a weekly news magazine to a well-researched autobiography, or perhaps a photo-journal covering an entire decade?

The web would work if we could see all its pages and all its sites at all at once.

Unfortunately we open windows, we click on links and something occurs, but without the guiding eye of a librarian, or the simple insights gained by a rudimentary system of filing, we surf, unaware as to the depth of information governing our progress, or without a clue as to the data currents we may be following.

Clearly there is a need for some form of hypermedia referencing on the web, (which today would be the Semantic Web) perhaps a visual means by which we may keep track of our chance discoveries, at the same time allowing for the creation of new maps, archives and desktops so that a patina is created over the worlds knowledge, which is not simply over-written by the constant need to reformat web sites using the latest craze or fashion.

Information must be allowed to decay within cyberspace and to exist in a form of electronic harmony with whatever may be occurring in the outside world today — it makes no sense to constantly refashion everything only to find that there is no possibility of a return to stable, old-fashioned content.

The opportunity exists therefore, to grow the web alongside our memories, to create personal histories and to share them, hence the need for Coleridge, a map-making project  successor to Xanadu, and the Memex from whence it sprung.

The Memex was first described by Vannevar Bush in an article published in Atlantic Monthly magazine in 1945 entitled “As we may think”, and describes a desktop apparatus comprised of a slanting translucent screen “on which material can be projected for convenient reading” including a keyboard, and sets of buttons and levers which would enable “trails” between different and varying texts.

The Memex idea, along with card filing systems, has been around in one form or another for over half-a century. The question then is how do we organise our personal libraries, our personal filing systems to the global brain over the next half-century, without losing history?

One could simply enforce the attitudes of a strict librarian and resort to an empirical system of cataloguing. However, like Dewey, I believe the web to be similar to a disordered pack of cards.

Each enquiry produces a new order, or an organic whole. In fact the underlying order is less important than the significance or sense made from each foray into cyberspace.

But this sense is changing as patterns emerge. In some parts, there is order, in other parts, nothing but dark space — the vast sea of the unknown. Unfortunately, and in spite of what everyone says and does, the web is a lot less present and localised than it used to be; less a place of discovery than a realm of the vast undiscovered, a seemingly foreign, territory.

At the same time, a form of calcification and stagnation is occurring as we stick to our backyards, congregating around the same portals, like Yahoo or Hotmail, while the vast sea of unknown, uncharted information grows and grows. One could also liken this process to the creation of a reefs and islands, (or the clusters and nebulas from whence stars shine)

Invariably, if lost in cyberspace, one resorts to strategies from childhood —  you ask for directions, a neighbour points the way, or a search engine directs you to a web site.

Surfing on the other hand is still a risk, since any link may veer off in the opposite direction, costing you time and time is money.

What we have created on the web is essentially a benign form of global chaos in which there exists small zones of order, a lot of which is nothing more than an ever-increasing popular  market.

The reverse should be true, what we should be experiencing are huge wave or movements of individual pattern-making as the chaotic eddies pan out, running past the fast rapids of (to use a geographical metaphor) a young mountain river, and tending towards a slow Nile or Amazon basin in which all forms of life occur.

There is a demand then, for a visual reference point,  organisational clues, perhaps even a generalised representation of the medium like the globe or atlas with its lines of latitude or the meta-links which separate one form of hypermedia from another.

The dynamism of the web would disappear if we imposed an artificial order that was too strict, and likewise if we favour one or two portals over another we lose what is  individual and public/private to a mall, a  commercial market.

Still there is a need for a visual reference, a way of conceiving the web beyond merely a rough collection of sites and pages.

The need becomes more apparent, the more one realises what most ordinary people fail to see — what they are looking at has no real structure except for a form of discontinuity — one page is seemingly as important as the next; information is called up or falls away; and so the public is at a loss to explain what is occuring when we surf.

Inundated with sensory stimulus, we resort to the metaphor of television and the visual world which is part of broadband media. The reverse should be true. Cyberspace should exist as a pleasant, alternative experience in which we travel, walk, move about and talk.

What we could be doing then, is promoting a means of visual referencing perhaps sharing our personal desktop adventures, without risking our privacy or losing the invisible line separating what is personal and private from what is public.

Separating commercial web-sites from public and private web pages and creating a sense of global village responsibility along with boundaries and borders might do the trick.

Simply creating a well sign-posted highway and providing guide books and maps is another solution.


The better solution though is to allow everyone to participate in the map-making process, and I believe this is best achieved through a form of visual transfer of information such as Coleridge may offer.

Clearly there is an ancillary danger that our individual myths, our conceptual boxes or intellectual packets will be lost, as everything tends towards one dominant idea, for instance the Web is just such an idea, as is the Net.

There is no reason though, why cyberspace should look like a tangled spiders web, or why the Net instead of being a fine Mesh, is more like a Maze . Is there something wrong with the basic structure of cyberspace? We forget that what started out as a few pages has turned into entire sites which have grown into portals and so on, there is a type of order, but this order is fractured. A sense of wholeness may arrive, but only if it is allowed to exist alongside our own map-making processes.

Encouraging our quantum, personalised contributions to the organic whole-making that is apparant on the Web, is one approach, and there are a number of ways in which we may regain our sense of purpose in creating personalised hyperlinks and the individual exchange of information.

Coleridge is possibly a first step in the creation of a unified hypermedia whole in which the individual mind still exists alongside the machine. It is also a plea for an open, non-commercial response to Xanadu which I envisage as some form of simple tool or applet, even a desktop, that could sit conveniently alongside and interface with current browsers and hard-drives.

Coleridge as the Model of Hypermedia Organisation

some first steps.

1. A *Navigator or *Explorer window which contains nothing more than route maps and visual outlines of a web site, a web system or internet galaxy.

2. Internet galaxies could contain a series or set of web sites. Each site is really a system of interlinking pages and hyperlinked sites which form a whole. Each system is characterised by the type of content to be found. On further exploration, these systems might turn out to contain planets or perhaps a reef surrounded by small islands, but the journey is always personal and pleasurable.

3. The creation of a uniform system of meta-tags and embedded icons which would enable the next generation of browser to navigate using various maps. A map server or search engine might provide additional paths and extra tags/icons. Eventually we would download our Coleridge maps and upload our own as a form of sharing coexisted alongside more commercial attempts at advertising and marketing the existence of information and so on.

4. An important, visual distinction would have to be made between commercial sites, public resources, organisations and individual home pages. Instead of lists of addresses, one would find strings of connected icons, each one indicating a feature or quality about the information to be found. Maps and icons would be nested allowing levels and a form of order to coexist. Perhaps some sites would appear as trademark icons, while others were marked unknown, like an uncharted island. Or perhaps a personalised icon might tempt us to create avatars, as we popularised our home pages within the public sphere? The important thing is that we don’t lose the personal (and the political) in the plethora of commercial intrigue that populates the current electronic universe.

A mind map, showing visual as well as semantic information

5. Clicking on any one of the icons in the Coleridge window would immediately transport you to the site or system, however, like the small maps found on some CD Roms, a single click would return you to wherever you started, or allow you to jump back and forward at warp speed. Gradually, users would create their own Coleridge maps, or swop them as cool Coleridge files, as a unique, individual sense of discovery re-occurred.

6. One can invision a Coleridge map taking on the personality or quality of its information, as users create new icons, and window backgrounds. As a general rule, the exploration enabled by Coleridge would be fun and user friendly. If maps are too large, then they will become cumbersome, however, an academic or commercial project might wish to create a library or single archive based upon Coleridge, a place in which all points on the web are easily traversible, from one central mapping facility.

7. Essentially, Coleridge has been available for some time — whenever we happen to stumble upon a directory,we find files along with their icons and if we are lucky, their names, dates and file sizes. Coleridge would give this kind of information greater meaning, while taking the directory scanning process up a notch or two. Instead of files, we would scan over web sites, instead of web sites, constellations and galaxies of new media.

*Whether these programmes or more traditional operating systems should be the ones to patch us into the Coleridge interface is open to debate, but from here, it seems that any one, or all of the above may be true, and a degree of novelty may occur in how each solves the dilemma of competing against an open system. A note on the public domain Since Coleridge is really another desktop interface with the web, what we are creating is a form of spontaneous visual communication, a subjective mapping and modeling of a global system. There is no sense in putting up yet another department store, or commercial interface right outside ones front door. When we interface with the real world, even if the real world is virtual, some semblance of freedom and liberty should occur.

The Coleridge experience is a way of moving out from home pages, into streets, valleys and open plateaus – whatever version of the landscape turns you on. The project is therefore awarded to the public domain, so that no human (or android) gets to robot over us, and a form of virtual community or global village is allowed to exist within the jungle. Get the picture?

May the best version of Coleridge win! Long live Xanadu

Also, keep on surfing and remember cyberspace is the one place where the map is the territory.

David Robert Lewis Cape Town, Saturday 6-1/2001

PO Box 4398 Cape Town 8000 Republic of South Africa

Leave a Reply