It is clear that the World Wide Web and Hypertext has its roots in several pursuits most notably those outlined by H. G. Wells (1938), Vannevar Bush (1945), Joseph Licklider (1960, 1963), Theodore Nelson (1963, 1999) and Douglas Englebart (1965). However, it is also clear that the original visions of how the world’s information might be organized and communicated electronically via hypertext still have not materialized (see Nielsen, 1995, for a brief history of hypertext ideas and developments from Bush to the early WWW). For example, Bush proposed associative knowledge trails by way of his memex machine (1945). Nelson proposed hypermedia unified by transclusion (1995, 1999) by way of his xanalogical structures. However, the state of current hypertext on the WWW does not include the affordances required in either of these visions. Why?
This project is my attempt to answer this question. This pursuit began with developing an understanding of the history of how various hypertext ideas and systems were developed. That was accomplished via a review of the literature cited in the previous paragraph. Then, in order to better understand Bush’s specific vision of associative knowledge trails and side trails, I decided to create some myself that took advantage of current affordances but still attempted to maintain Bush’s original vision.
After the literature review and associative knowledge trail learning activity, I felt I had enough information to create a hypothesis to explain why the more complex hypertext systems posited by Bush and Nelson are not commonplace now. I make conclusions and suggestions for further research and inquiry.
The associative knowledge trail entitled Roots of Hypertext (described below) is the container for my literature review. I think it successfully captures not only the literature reviewed for this project but also the connections between resources that I made while reviewing them.
My Process of Creating Associative Knowledge Trails
Bush (1945) envisioned a system whereby the user creates trails and side trails connected to a particular interest. As described, it would appear to be a linear progression of resource associations with the occasional tangential side trail stemming from the main trail and the occasional annotation added by the author of the trail.
The first step was to decide upon a digital container for my associative knowledge trail (which I acronymed AKT in my notes). I evaluated different online platforms to ensure that the affordances I needed were included. Here is the list of the affordances I was looking for:
- hyperlinks can be created to online and offline content
- a method for containing the structure of the trail itself (e.g., timelines, flowchart, arrows, etc.)
- image and text objects can be embedded
- a method for logically showing and presenting side trails is included
- a quick and easy method for adding annotations
- generally, app must be is easy to use, and quick to develop the trail
- a method by which the trail can be shared online
I assessed ShiftSpace, online visual timeline creators (the best of which was Capzles), CoKnown, bubbl.us, VoiceThread, and Prezi. I also looked at local computer applications such as SMART Ideas, SMART notebook and Microsoft PowerPoint.
As much as I wanted an online, Web 2.0 app that I could share as part of this project, none of the online tools I examined provided me with all of the affordances in my list. Capzles came the closest but it did not have the ability to create links to online documents or video (all multimedia content in Capzles had to be uploaded). I decided to use SMART Ideas. This program can export all layers as HTML so it was possible to share my work online without the need for the original application.
The next step was to begin the creation of my AKT and keep a record of my reflections as I created it. I decided to create an associative knowledge trail entitled Roots of Hypertext. (If you have SMART Ideas installed, here is the original IPR file.) I also wanted the AKT to serve as my literature review. I gave myself a limit of three evenings to create my the AKT. This included reading and reviewing resources and integrating them into the trail. In the process, I reflected upon a number of ideas and issues that arose during the activity.
Summary of Reflections
(See Appendix I for all reflective notes)
I learned very quickly that my AKT was not going to look linear. The more resources I connected to the AKT, the more interconnections I saw, and the more connections I tried to make. Secondly, as the connections and layers grew, the AKT started looking more like a concept map than a knowledge trail. I started to categorize resources and colour-code them. I tried to make the AKT as visual as possible; a text based trail would have worked but I wanted to take advantage of visual affordances so that the AKT was more usable to me and welcoming to others. I also started to experience diminishing returns; as I added more resources to the AKT, it took longer and longer for me to make the connections between the existing trail and the new resource to my satisfaction.
If I continued with my SMART Ideas-based AKT, I envisioned that it would develop into a hierarchical organization of resources. This notion quickly reminded me of early attempts to organize all the information on the WWW on websites such as Yahoo and dmoz. These early efforts resulted in hierarchical categories of information. This sort of organization worked until there was too much information in the structure. Once the amount of content online became too great, hierarchical indexes of that information were impractical. Search engines such as AltaVista, OpenText, and Hotbot started to become popular because they pointed users to the information they were looking for very quickly. Google eventually perfected the search engine tool.
I considered my initial AKT creation activity using SMART Ideas to be a failure. I could not create an associative knowledge trail in the way that Bush had outlined. Nevertheless, I decided to try again and try harder to break away as much as possible from the concept map affordances of SMART Ideas, and really try to create a trail of knowledge with side trails created as needed.
For my second attempt, I did two things differently. First, I started with a brand new topic that I knew very little about and formulated an inquiry question rather than just a topic declaration. I have recently revived an old interest in electronics, so I decided to ask: How do AM radios work? (If you have SMART Ideas installed, here is the original IPR file.) Second, I decided to use of the layering feature as the method for my trail. That is, instead of trying to visually capture my trail on a single page, as I did in my first attempt (with side trails being the layers), my second attempt would use a different layer for each new associative resource. So, the path of my trail will travel through layers in SMART Ideas, not follow connecting arrows between symbols.
My hypothesis is an attempt to explain why more complex hypertext affordances have not been made part of the current web. My hypothesis is:
Information shared for consumption on the web must be organized in such a way that ensures the greatest accessibility and usability for the end user. If it is not, then the information will be largely ignored.
Hypertext, as it is used currently in digital documents, associates information that is related but does not make impositions on that information. The associative knowledge trail of Bush (1945) and the xanalogical structure Nelson (1999) proposes encumber the information they seek to present with an idiosyncratic or a needlessly complex organization.
Usability research examining hypertext on the web would seem to support this hypothesis. Jakob Nielsen (1989) concluded that “there is little hope for a single, universal hypertext user interface design which will be optimal to everybody” (p. 246). This would suggest that due to the varying needs of end users, the simplest hypertext interface will probably be the one adopted and used by the masses. Additionally, Nielsen’s central hypothesis in Designing Web Usability: The Practice of Simplicity (2000) is that simplicity ensures usability. In his characterization of the end user, he reveals what he believes is the underlying rationale for this goal: “People are extremely goal-driven on the Web. They have something specific they want to do, and they don’t tolerate anything standing between them and their goal.” (p. 380)
In the case of the associative knowledge trails, this form of presentation reduces the accessibility of the information because it tends to impose another individual’s personal organization of the information, one that is tied that person’s own abstract knowledge structures. My attempts at creating associative knowledge trails were not that personal because I could not push from my mind that reality that other people would be seeing it and using it. Perhaps, I am so firmly entrenched in the current social media ethos that I cannot create a truly personal knowledge trail.
A simple hypertext system facilitates web page usability for users. It would seem that any attempt at remediating hypertext systems to capture deeper associations or transclusions, will not be embraced by the populace as a viable information system because these knowledge associations are idiosyncratic and threaten simplicity of design.
The learning activities and literature review in this project have led me to propose this hypothesis. A direction for future research would be to test it directly with end users as subjects. Perhaps an experiment could be designed in which the information content is controlled but presented using different complexities of hypertext and differing levels of idiosyncratic organization of the knowledge. Subjects could be given a task whereby online information must be collected for some purpose. Time to complete the task, keyboard and mouse logs, and personal ratings/reflections/interviews could be potential data vectors for the study.
Reflections on creating an associative knowledge trails (AKT) for hypertext in SMART Ideas:
- My choice of hypertext was perhaps too complicated. There is too much information to be comprehensive. My AKT is, therefore, very selective and arbitrary and quite personalized as it reflects my own choices of resources used to understand the topic.
- Perhaps AKTs to be useful need to be highly focused, specific rather than broad.
- Perhaps the reason hypertext has been the most successful remediation, above AKT or
- The ability to add visual information, while more attractive to the eye and easier on the brain in terms of helping me to quickly know which resource was which, it took more time to add those visuals
- I began to colour code my entries and made a legend. I quickly realized that I could not continue very far with the method as I was repeating pieces of information. Bush’s original idea of an AKT is far more linear but SMART ideas, with connections and multiple layers, surely has affordances built in to allow for a greater complexity in the knowledge trail.
- Diminishing returns. As I added more and more pieces to my AKT, it was taking more and more time to organize and re-organize my resources.
- I keep thinking about my audience. I am making this not only for my own reference but also as a reference for other people. Can I do both?
- Do I organize my trail as a trail of ideas, a trail of resources, or a trail of authors?
- If I organize by ideas, my AKT becomes more of a concept map, which is fine, but not what I am looking to create
- If I organize by resources, my AKT approaches more closely what Bush was proposing in his Atlantic Monthly article.
- If I organize by author, it becomes more of a chronological reference list. This was my first instinct actually.
- Beginning to question how useful an AKT would be, if done exactly in the way Bush would have wanted it. Bush’s AKT conception was generated within the context of information in 1945.
- Thinking about the early days of the Web and how hierarchical indexes such as Yahoo and dmoz tried to organize all the resources online. Once the amount of content online became too great, a hierarchical index is impractical. Search engines such as AltaVista, Open Text, and Hotbot were early successful search engines before Google.
Bush, V. (1945). As We May Think. Atlantic Monthly. Retrieved from http://www.theatlantic.com/unbound/flashbks/computer/bushf.htm
Englebart, Douglas. (1963). “A conceptual framework for the augmentation of man’s intellect.” In Hawerton, P.W. and Weeks, D.C. (Eds), Vistas in information handling, Volume I: The augmentation of man’s intellect by machine. Washington, DC: Spartan Books. Retrieved from http://www.dougengelbart.org/pubs/papers/scanned/Doug_Engelbart-AugmentingHumanIntellect.pdf
Hendricks, V. (2005). Mainstream and Formal Epistemology. New York: Cambridge University Press.
Licklider, J. C. R. (1960). Man-computer symbiosis. Human Factors in Electronics, IRE Transactions, 1(1), pp. 4-11. doi: 10.1109/THFE2.1960.4503259
Licklider, J. C. R. (23 April, 1963). “Topics for Discussion at the Forthcoming Meeting, Memorandum For: Members and Affiliates of the Intergalactic Computer Network”. Washington, D.C.: Advanced Research Projects Agency. Retrieved from http://www.kurzweilai.net/memorandum-for-members-and-affiliates-of-the-intergalactic-computer-network
Nielsen, J. (1989). The matters that really matter for hypertext usability. In Proceedings of Hypertext ’89, (pp. 239-248). Pittsburgh, PA: ACM.
Nielsen, J. (1995). Multimedia and hypertext: The Internet and beyond. Cambridge, MA: Academic Press.
Nielsen, J. (2000). Designing web usability: The practice of simplicity. Indianapolis, IN: New Riders Publishing.
Nelson, T. H. (1965). The hypertext. Proceedings of the World Documentation Federation.
Nelson, T. H. (1995). The heart of connection: hypermedia unifed by transclusion. Communications of the ACM, 38(8), 31-33.
Nelson, T. H. (1999). Xanalogical Structure, Needed Now More than Ever: Parallel Documents, Deep Links to Content, Deep Versioning, and Deep Re-Use. Computing Surveys 31(4). Retrieved from http://www.cs.brown.edu/memex/ACM_HypertextTestbed/papers/60.html
Wells, H. G. (1938). World Brain. London: Methuen. (Also available from http://ebooks.adelaide.edu.au/w/wells/hg/world_brain/complete.html)