The Paris Pages Experience Index 
Discussion
In the coming years, all the the issues mentioned thus far will receive significant attention. Yahoo!, Lycos, and others are spending millions of dollars right now on how to classify diverse data types effeciently. Perhaps many of these approaches will remain proprietary for a while, but perhaps not.
In summary,
- Classifying/Characterising diverse data types is a fundamental problem in web site design and presentation.
It is essentially a database definition and database construction problem; and it is a difficult problem
given the diverse sorts of data that we wish to present in a web site.
- Associating different data types/categories/classifications is a fundamental web site planning activity.
It is also a difficult problem because in general there is an infinite number of ways diverse data can be presented / arranged.
Fixed and static html files are not easily rearranged, and certainly not in real time.
Dynamic html coding is difficult. At this time a combination of fixed html files, and real time html
creation may provide a rich hybrid approach.
- Hyperlinking data can be complex, but it is essentially the creation of the web site itself. Thankfully the number of links does not in general grow exponentially, but rather is linear in the number of html pages. There are fewer in a hierarchal presentation (mostly vertical linking) than in a relational presentation. (horizontal linking allowed). At this time creating rules to decide whether to link, or not to link, is essentially a human activity, although computers can carry out the linking process. Thus highly complex linking rules remain problematic.