This is a report of an open lecture given by Professor Ted Nelson and organised by Ron Burns, Department of Engineering and Computing, Bournemouth University, UK - 6th February 2002
Ideas, the Final Frontier - Computers Beyond Hierarchy and the Web Beyond HTML
Ted Nelson began his lecture by recalling the standardisation of rail track widths in the 19th Century (when Charles Babbage was head of the Royal Commission into the rail network), as well as the disputes which occured in the 20th Century over the use of the radio frequencies AM and FM. New standardisation wars are now 'flowering' in the computer field, he said.
He then spoke about the computer as a protean machine, used for the storage and mapping of information, and concerned with complexity. He asserted that the need to 'respect the subtleties and complexities of each area' and type of information was often overlooked by 'geeks' who 'take control of the field'. The 'geeks and techies' are ignorant of culture, don't understand human life, thoughts and feelings, and the aggressive ones try to impose their ideas through their software applications.
(Games software is better designed than office software, he said, because those who design games software like playing the games, where as the office software designers mostly don't enjoy using it.)
Nelson doesn't like the way computers have been engineered to imitate paper; and the way that the hierarchical structuring of data is standard, and he said that 'computers are hierarchical because we make them that way'. He then went on to talk about the World Wide Web, which, after 7-8 years of existence, is presently held together by legal arrangements between Microsoft and the WWW Consortium. The internet was a flat space of names until the URL was invented, and the related path names and directories have created the hierarchical structure of the web. Ted Nelson doesn't believe it will stay that way, and regards the current web as a 'dumbing down of ideas'.
The GUI created one kind of standardized interface: the desktop with file icons, while the 'application' program is like a prison, usually preventing the activation of any data in any other program. The creator of the code therefore almost owns the data - made with 'Paint' opened only with 'Paint'. This has removed the opportunity to program from a lot of users and left the programming to the 'druids'. Corporations like Macromedia and Adobe, who own many of the graphics and multimedia applications, fight to maintain their file formats. In contrast to this he then mentioned open source code, GNU Linux for example, but said he thought there needed to be a balance between corporate control and open source code, and that 'it can't all be open source'.
"What is structure?" Nelson asked, and then compared the engineer's typical view of structure as hierarchical, while artists and academics were aware of more 'gentle' structures. He said that the latter know that things change and that they need to keep track of transformation. They therefore need a different structure. Our imaginary construct, or how we see the data, very often doesn't match the view of the corporation who designed the application or 'imaginary construct' which we agree to use.
Ted Nelson then showed his own software 'ZigZag', which has a visible skeletal structure which is: 'A different way of thinking about the universe of structure'. A different 'cosmology'. For one thing, he wants to make visible that which is not visualizable on paper. He also wants to 'set the cells of data free' within a single unified environment.
His early background was in the arts - some might say he was a protean new media artist - he wrote a play, did some acting, wrote a book, a rock musical and made a film. This experience showed him that to make these works required a layering of processes and a use of mechanisms like 'travelling mattes' to create special effects in film, or four colour printing, or layered audio (which was just beginning in the 50's). At graduate school, he did a computer course and this was how he came to invent hypertext, and Xanadu and foresee the development of the PC. Xanadu was bought by AutoDesk, but has so far not been developed successfully. The WWW uses hypertext however, but as developed by Mark Andriessen (Mosaic) and Tim Berners-Lee (Netscape), the WWW has a directory based structure and URL based protocol. Nelson explained that his concept of hypertext was all about content. Specifically, it was about comparing different versions of content in order to create a multi-dimensional context.
The WWW, and the browser which simulates a page, has, in Nelson's view, no content management technique. It is trapped in a prison of paper simulation, and hierarchical protocol - 'the trivial model of directories'. For Nelson, XML - the next code for the Web being developed by Berners-Lee - is potentially a triumph of hierarchy over other content management possibillities. He wants nothing less than a re-evaluation of how we manage content, both on computers and on the Web. It needs to be re-usable and simultaneously comparable. He proposes 'transcopyright' as a way of creating a 'quotable universe', and he means big quotations rather than sound bites.
Nelson said that there are no pure technical issues, and that these are always complicated by issues of business and, more importantly, aesthetics. The issue of copyright, which is one of both business and aesthetics, has been ignored in his opinion, by the WWW Consortium, who merely see it as a purely technical matter. He talked about transcopyright and demonstrated diagrammatically how the concept would work. There would be a 'content-pool' of works by participating authors. These works would be for free distribution, in the possible format of VLIT files. This content could then be represented in any way that the user wants, within any interface. Payment could be made for access to the content-pool, or for using individual source documents.
'Everyone should have the interface they want' he said. The important thing was to keep all the contexts, and to be able to relate contexts side by side. However, this is only possible if the user can correct any misinterpretation, derogatory or otherwise, if they can return by one link to the source document. The ability to see the evolution of documents through the re-assembled content would potentially be a fair and transparent way of managing content. The current system places too much emphasis on flashy and instant information rather than understanding through deep contextualisation.
ZigZag, meanwhile, is a multi-dimensional 'skeletal structure for representing the atoms of relationships'. It is non-hierarchical with 'profuse overlap' - the complex overlapping of indexes and content. He said that categories are entirely viewpoint dependent, and that he wanted to see manifolds of profuse relationships - the structures of ideas and assertions.
When asked a question about possible visualisation of this n-dimensional space, he dismissed the idea of 3-D Euclidean space, and instead spoke about the paradoxical variations of space. Visualizing a multi-dimensional space to an individual's viewpoint would be possible if the points on the mesh of this n-D mapping could be used relatively.
In answer to questions about content, meaning and structural relationships, and criticizing the Semantic Web as a 'geekly proposal' designed to map meaning, he said that there are thousands of ways to describe anything, and that any mapping is deficient. It is the search for the philosopher's stone, as ever.