The original vision of Vannevar Bush, Joseph Licklider, Ted Nelson, Doug Engelbart, and Alan Kay is a computer that supports the user in cognitive processes – a personal dynamic medium for creative thought. In variation of Isaac Asimov’s first law of robotics** Isaac Asimov’s first law of robotics is «A robot shall not harm a human, or, through inaction, allow a human to come to harm.» The law is taken from Asimov’s science fiction novel I Robot (Bantam Books, New York, 1977).
[Update 7/2016: The Three Laws of Robotics have been introduced 1942, cf. Wikipedia: Three Laws of Robotics], Jef Raskin writes forty years after Licklider’s article on Man-Computer Symbiosis the first law of interface design. It reads [Raskin 2000, p. 34]:
Any system shall not harm your content or, through inaction, allow your content to come to harm.
The necessity to formulate such a law demonstrates how far away we are from the original visions.
The present thesis has shown that the foundations of hypertext are already rooted in the 1960s. Doug Engelbart, Andries van Dam, and Ted Nelson, implemented the first systems that are capable of hyperlinking. The 1970s were relatively quiet. But with the advent of personal computers in the 1980s, a lot of new hypertext programs were developed. They utilized WIMP concepts on the Apple Macintosh, on powerful workstations, and later on Microsoft Windows for IBM-PCs.
The foundations for the present form of graphical user interfaces were in the 1960s and 1970s. This thesis recalls the invention of the mouse by Doug Engelbart and Bill English, overlapping windows and popup menus by Alan Kay, and icons by David Canfield Smith. The parts were assembled in a consistent way for the Xerox Star in the second half of the 1970s. Extensive user studies, task analysis, and applied cognitive psychology produced evidence for the development of the desktop metaphor. These efforts can be rated as the first emergence of user-centered design. The industry needed a few attempts until the Apple Macintosh defined the standard for graphical user interfaces as we know them today. Since the late 1980s, the pace of innovation on the field of user interfaces for pcs came to a halt. The section on filing systems has shown that the dominating operating systems with graphical user interfaces do not offer sufficient methods for the user to cope with thousands of objects (cf. 3.4.1 Filing). The concept of direct manipulation and the paradigm of the desktop metaphor do not scale to the vast amount of items we are managing today. They were right in the beginning – today they are inadequate. Innovation is necessary to regain control of our files.
The situation became even more complicated with the tremendous success of the World Wide Web. The user interface for the Web, that got momentum in the early 1990s, unfortunately lacks a profound approach of user-centered design. No one felt responsible to start an attempt comparable with that of Xerox Parc’s research for the Star computer or Apple’s research for Lisa and Macintosh. As a result the Web interface was never integrated into the desktop environment.
A solution can be found by reflecting upon the core values of hypertext and graphical user interfaces. The following considerations are in accord with the present thesis.
The fundamental idea of hypertext is the relation between different texts. The common means to express such structures are hyperlinks. In As We Should Have Thought [Nürnberg/Leggett/Schneider 97] – another pun on the title of Vannevar Bush’s article As We May Think – Peter Nürnberg, John Leggett, and Erich Schneider criticize linking for two reasons [Ibid., p. 1]:
Firstly, linking implies a certain kind of structural paradigm, one in which the user […] links information together for purpose of navigation. […]
Secondly, linking implies the primacy of data, not structure.
Navigation is important, but it should not be the only purpose of hypertextual structures. Especially section 3.4 Provisions for the Future of the Desktop Model (p. 86) has shown the need to express various kinds of relations in a flexible way. Neither hierarchical nor navigational structure should be imposed on the data.
The second argument opens the discussion to put structure, opposed to data, into the center of computing. At least, structure should not be considered as a second-order attribute to data.
Graphical user interfaces have the potential to convey such structures. Humans are able to understand abstract concepts, that go beyond the desktop metaphor. With Ted Nelson’s words [Nelson 97a, p. 21]: «The fundamental information problem is to keep track of ideas, and represent them, accurately.»