The main reason why Hypertext has a fascination on me since a long time, is probably the fact that it acts as a bridge between two basic, opposite, and complementing elements that may be called gender of knowledge representation: tree and shortcut. Where shortcut is meant in a broader sense, including genuine cross-references (“see also” as opposed to “see”), genuine shortcuts that jump between different hierarchy levels, and direct access links from a virtual top level, like aphabetically arranged articles and search engine hits.
The opposition involved here has been the object of many endless religious/preferences/taste wars. In the paper era, it was mainly about systematic classification catalogues vs. index words catalogues, or systematically arranged encyclopaedias vs. alphabetically ordered dictionaries, while today it’s about browse vs. search, or about deeper website hierarchies with shorter pages and longer click paths, vs. longer pages with shorter click paths but longer scroll distances, and about orientation need with serendipity effects vs. fast access desire. In the paper era, only one of either approach could be implemented, and a decision had to be made.
For genuine network structures, however, both approaches are inappropriate. Consider, for instance, a dictionary on (German) principalities’ territorial history, or on etymology. Alphabetical order is very unsatisfactory here, because there are so many types of connections hidden: in addition to the chronological relationships, the geographical and dynastical links in the former example, or the morphological and the semantical relationships in the latter. Chopping the content up into alphabetically arranged atoms, lets the user lose orientation. However, the tree structures showing through are not strict enough to be applied as the main ordering criterion.
Hypertext can well handle both, trees and networks, by combining hierarchical links and genuine cross reference links leading from one branch to another. I like two examples showing a hypertext-like structure without actually being a hypertext.
- The first is the Dewey Decimal Classification used by librarians. At first sight, it looks rather strange to try and line up all things in the world on a linear axis of numbers. Looking closer, however, a sophisticated network of cross-references can be discovered, which includes borroughing sequences of digits from another branch with related meanings, or giving demarcation hints in the section headers or in the manual.
- Another sample is the (German) telephone numbering scheme. If you follow the path of a connection being switched between 06222 Wiesloch and 033932 Fehrbellin, first upstream to 069 Frankfurt/M. and then downstream from 030 Berlin, you might wonder whether the connection to the neighbor village 07253 Bad Schönborn might use a similar hierarchical path which would be a big detour. Of course not, there is a shorter cross link.
In both examples, the intelligence of the system resides in the smart combination of trees and shortcuts.
Pure tree is dogged, pure network is chaos; knowledge is hidden in the combination of both; Hypertext has the potential to implement this bridge. Speculating one step further, it could, in the long run, even bridge different cognitive styles, and perhaps also the science/humanities divide related to these different styles.