Wednesday, September 23, 2009

Theodor Holm Nelson: “Most people are fools, most authority is malignant, God does not exist, and everything is wrong”

by David F. Bello

Ted Nelson’s biggest effect on the universe of computing is through his disdain for the limitation of specific cultural models on its development. Through his fifty-year-long failure to engage widespread adoption of his own model of hypertext, Project Xanadu, he bitterly declaims the restrictive qualities of technology as we use it, and longingly pines for the breaking down of conventional modes of metaphorical representation.

In other words, it doesn’t have to be this way; it can be any way. We must utilize the best way.

When I read Geeks Bearing Gifts last year, it revealed to me a revisionist perspective on computer history. Nelson is to the history of digital systems what Howard Zinn is to American history.

The cover of the book is a mugshot of Bill Gates with the subtitle, “How the computer world got this way.”

Is there a “computer world” we can talk about so generally, and If so, in what way is this computer world? Why does Nelson see this as a bad thing, and not merely the grand, consistent meta-narrative of functional technology? Finally, what does he propose to do in order to change it?


Yes, a “computer world.” In proposing ZigZag, an alternate “computer world,” Nelson speaks of a cosmology, or “a particular account or system of the universe and its laws” (via, requires university access). In this sense, the “world” of computing is the cultural conception of what computing
is, how and why it works, and overarching ideas about what is can and cannot do or be. Can this be more general? Probably not; and that is Nelson’s point: everything we take to be concrete, immutable, inevitable, permanent, and inflexible about computing is only a single branch of potential ways of being!
I propose to go back to the beginning and branch in another direction. Let's go back to 1945, say, when anything seemed possible, and when conventional wisdom could be challenged more easily because it had only been made up a few months before, perhaps by another 20-year-old in the same lab. The evolving conventions of those days involved lump files with short names, hierarchical directories for storing them, programs as long strings, databases as tables stored in files. (These would later be packaged into "folders" and "applications".) But now imagine that a few guys had snuck off to another lab and created a whole new system of conventions, completely different, for everything (via).


The computing world is based on one principal system of conventions -- the simulation of hierarchy and the simulation of paper (via).

Because computation has the relatively unique capacity to interact with information recursively, a system of data representation that acts recursively was adopted in the hierarchical file system. Folders can have folders inside of them, but what are they other than lumps of files? Nelson desires a more fluid and multidimensional model for storing information, which he attempted to instantiate in Project Xanadu, but could not compete against the ideologically adhered to boundaries of the traditional computing industries of Microsoft, Apple, Xerox, Adobe, etc.

(On a somewhat tangential sidenote, in his book Gödel, Escher, Bach: An Eternal Golden Braid, Douglas Hofstadter describes human thought in terms of “strange loops” and “tangled hierarchies” which do not conform to typical representations of data structures and interpretation. If we were to eventually create a computational model of human cognition, it would most certainly not be confined to a strictly hierarchical model.)

Whether you are happy with hierarchy being imposed on your day-to-day computer life (along with the idea that it is imposed on nearly all computer use as well, in some form or another) or not, Nelson’s primary point is that it should not be imposed at all. There should be multiple branches of computer history, starting back at that point in the development of traditional file system architecture. So, when we talk about current “filesystems” in common data storage, we are really talking about multiple instances of the filesystem.

This idea is more apparent in Nelson’s discussion of the PARC user interface (or PUI) in the video at the start of this essay. Every graphical user interface used in PCs today utilizes the desktop metaphor, some kind of taskbar, shortcuts, a buffer folder of files ready to be deleted (such as the unmodifiable “trash” or “recycle bin” folders), a metaphorical “clipboard,” a two-dimensional plane where the mouse can “act,” “shortcuts,” “windows,” and software applications. However, there are many other ways of envisioning a graphical user interface. These are simply the conventions adopted over the course of computer history. These conventions make up, in part, what he calls the cosmology of computing, along with the simulation of paper.


One of Jakob Nielson’s heuristics of user interface design is as follows:
Match between system and the real world. The system should speak the users' language, with words, phrases and concepts familiar to the user, rather than system-oriented terms. Follow real-world conventions, making information appear in a natural and logical order (via).

Nelson would see this as a grave restriction on the potential for data representation models. Of course, the learning curve for something like this would be much shorter because you already have knowledge of an existing cognitive system that has been mapped using digital metaphors of that system. However, we are losing out on the possibility of digital representation to come up with a literally infinite set of computational modelling systems by exclusively adopting this cosmology of hierarchy and paper simulation. Metaphors make this simpler to understand, but when they are adhered to at a fundamental level, they make it difficult to innovate outside of their bounds.

One simple failure of the way we simulate paper using computers comes from the idea of “cutting” a piece of text and “pasting” it somewhere else among text. Nelson describes the failure in Geeks Bearing Gifts:

The term "cut and paste", as used by writers and editors for many years, refers to rearrangement of paper manuscripts by actually cutting them and physically rearranging them on desktop or floor. It is a process of parallel contemplation and rearrangement, where you look at all the parts, move pieces around, put them in temporary nonsequential arrangements, and choose a final sequence. In 1984 this changed: when the Macintosh comes out, they change the meaning of "cut" to *hide* and the meaning of "paste" to *plug*. To be fair, many people, not just the PARCies, imagine that the serial process of hiding and plugging contents is THE SAME as the parallel process of rearrangement. This original parallel rearrangement process is fundamental to writing and revision, especially in large-scale projects.  Because no good rearrangment software exists, it is far harder to organize large projects on the computer than it was with typewriters, before the PUI.  The organization of large projects has become far more difficult, unless you print out and physically cut and paste the old way (via).

So, we are forced to communicate our ideas in only two dimensions. Unless...

(Assuming you didn’t want to sit through the ten minute video above, Nelson demonstrates an alternate cosmology for data representation that acts in multiple dimensions of relation.) The two-dimensionality of common computer use as it is today exists due to the WYSIWYG paradigm of early document design software.


Nelson’s book, Geeks Bearing Gifts, tells the history of what he calls the “paperdigm,” or the conventional adherence to paper-simulation, as having come out of the PARC user interface. He claims that the reason they stuck to a two-dimensional model of design for documents came from the fact that they were funded by Xerox, who wanted to sell printers. Wanting to sell as many printers as they could, they forced the first and most influential user interface designers to incorporate printability as a fundamental characteristic of document creation. A cynical view of why our computing capabilities are so limited, but nonetheless entirely believable.

To look outside of this paradigm, we must consider the idea of hypertext, a term Nelson himself coined in 1963. If text is two-dimensional, hypertext is three-dimensional in that there are links to other documents existing in parallel. However, as Nelson envisioned it, he is incredibly disappointed with its misuse as a building block of HTML and the World Wide Web.

HTML is precisely what we were trying to PREVENT-- ever-breaking links, links going outward only, quotes you can't follow to their origins, no version management, no rights management (via).

In addition to these complaints, the WWW also imposes hierarchy in a number of ways: “the exposed hierarchies of the directories and pages; internal hierarchical markup (HTML, XML); and Cascading Style Sheets” (via). So, the beast that he created corrals him within the same adopted standards from which he was attempting to break away. Nelson’s bitterness is perhaps his most important characteristic.


Nelson's hatred of conventional structure made him difficult to educate. Bored and disgusted by school, he once plotted to stab his seventh-grade teacher with a sharpened screwdriver, but lost his nerve at the last minute and instead walked out of the classroom, never to return. On his long walk home, he came up with the four maxims that have guided his life: most people are fools, most authority is malignant, God does not exist, and everything is wrong. Nelson loves these maxims and repeats them often. They lead him to sympathize, in every discussion, with the rejected idea and the discounted option (via).

Nelson’s direct impact on the history of human-computer interaction involves the development of hypertext as a part of the WWW and its general use, but Nelson is displeased with its application. He tried to impact computer history with Xanadu and ZigZag, but neither have been widely adopted. His primary impact on the way we deal with and understand computing is not through specific implementation of key ideas in the past or present development of software or hardware systems. No, the most important result of his decades-long career is found in his dissenting attitude and optimistic proclamations of alternative cosmologies. Steve Wozniak says that he “was one of the most inspirational figures in the home computer revolution” (via). Wendy Hall, president of the ACM, predicts that Geeks Bearing Gifts will “be one of the most important, and potentially one of the most influential, computing books ever (also via). Stewart Brand also provides an advertising blurb for Nelson’s work alongside these most notable names in the development of computing.

As part of an internship I had as an undergrad, I attended a conference where Ted Nelson spoke. It was the first time I had encountered his ideas, outside of knowing him for having coined the term “hypertext.” In describing the ideas I’ve outlined above, Nelson blew my metaphorical mind when it came to computing. I assume he is able to do the same for many others. To realize the incredible, unending potential for digital representation and interaction was to forever become a proponent of far-reaching innovation without any restriction at any level.

No comments: