Have you ever articulated how you perform web research? This article focuses on how we might use, or mis-use, hyperlinks as we collect pages relevant to a topic of interest. We begin an exercise to explore “exponential growth” and its applications in viruses.
Vannevar Bush’s Memex Organizes Web Research.
A key concept of the Internet-based World Wide Web goes back to the influential article “As We May Think”, published in Atlantic Magazine in 1945. The government wartime science leader, Vannevar Bush, laid out a vision of knowledge work that would expand war-driven scientific advancements into decades of national science progress. He named the machine Memex (for Memory Extender) and sketched a brilliant, convincing scenario for its operations.
Bush envisioned a machine that organized documents to support thoughtful recorded and reusable “trails” of mental associations during a research project. But, something seems odd. The machine view of programs and apps feels familiar while the thought process remains an aspiration.
Our current networks are partway there. Behind HTML pages, the “semantic web” propels associations to a trail blazer’s attention. Separate text applications may record notes on associations and critiques and vocabulary and supporting calculations. But, a Memex trail is an object integrated with the researcher’s thought processes. Recovering thoughts and assuring integrity about a trail challenge modern web creation.
Aren’t’ we already living with Memex? Why would we need more thinking and recording? Can’t we associate and record as fast as we can type and speak? “Trails” appear cumbersome, or so we may believe without experience with structured hyperlinks. Yes, the WWW is flawed.
HTML Works Over “Loose Links”.
We’ve coined this term to question a fundamental mechanism that got the World Wide Web going. “Loose” means:
- one-direction, finding targets by searching semantic layers;
- not necessarily persistent, target URL content may change;
- used for navigation, ‘headers and columns and “read more”;
- arbitrarily named, relying on underlined words and surrounding context;
- lazily described, “so-and-so writes”, “here”;
- confused with novice instruction, “click here”;
- offered for favors, “my friend’s paper”;
- abused for profit, sold for behavior prediction;
- and inherently unstable, file systems change).
The semantic web adds subordinate layers with more structured references, but links remain “loose” in the minds of both HTML writer and browser clicker.
Bush’s article is a fantastic example of a “use case” in software engineering terms and for speculative prototyping. Its brilliant writing stimulates our imaginations.
Alas, Loose Links Scramble Thinking.
Here’s the odd part. Ignoring the massive technology differences between 1945 and 2020, we today lack the ability to record our thinking process and represent it in the hyper-mess of documents on the WWW. is this a much-needed gap? technology failure? human limitation? capitalist restriction? unimportant matter? correctable flaw? What is the value of a link?
How did the WWW fall short of trail as a higher level of linking thoughts and documents? Throughout the 1980s, a field of
hypertext research examined the value of links, while the information architecture field experimented with mapping hypertext into usable patterns for everyday navigation. Yet another field identified link labels that promoted issue-based information system design.
Did the WWW get off on the wrong design path? Sir Berners-Lee had enough to deal with transferring documents among file systems and other networks. A descriptive phrase and a file address seemed sufficient to one-way interlink documents to serve the work needs of the 1990s.
‘Loose links’ were great for jump-starting a world wide web. When graphical browsers came along, links could even be labeled with pictures that were descriptive to sighted people with screens. Since the human brain seeks interpretations of graphical objects, along came a way to blend graphics with beautiful typography to fit onto screens of all sizes, namely CSS (Cascading Style Sheets). At some point graphical objects gained an ALT attribute for description, left optional, to frustrate print disabled readers.
Anybody could write a link, name it in context or default to the unhelpful terms “click here” or “here”. So web page readers became accustomed to sloppy link conventions. Trails didn’t receive the support they deserved.
“Do links have morality?”
That is, “do link authors have responsibilities beyond offering a jump to a related document?”. Is hyper linking a philosophical matter?
“Yes”, David Weinberger opines about “The Linked Society”. Links demonstrate respect for readers. He recommends that: a hyperlink acknowledges a shared world between writer and reader, so the writer is obligated to offer some reason for injecting the link into the reader’s world. That is, the link seeks to expand their shared worlds and in ways that matter to both.
Skip to today, leaving links dangling, used for exposition as well as navigation, abused for surveillance capitalism, displaying un described graphics, and assumed as permanent fixtures of the now and future WWW.
A movement called web annotation seeks to provide the technology glue and intellectual stimulus to partially encourage trails, a later topic in this twURLED world.
Here’s my take on this75-year gap. The concept of a “trail” is neither necessary nor sufficient for modern personal web research. Yet, something is missing when I look at my saved search results, offline web pages, and aborted articles like this. My “twURLed World” struggles with haphazard link writing both inline and references.
The Memex trail is wishful thinking for a disciplined and durable web research process. But maybe I should try to change my habits by experimenting with my concept of a trail.
A Trail Experiment: “exponential growth”.
It’s easy to say that there’s something like rapid growth happening when a virus enters a population. “exponential” is a common variant , although somewhat vague in minds nod exposed to discrete mathematics. Let’s start with an image to encourage scribbling some calculations. My trail begins with a refresher parable:
Given a chess board, 8 by 8, add one grain of wheat in the upper left corner, then 2 in the next square to the right, continuing around the board doubling the number of grains square by square. How many gains will be in the last square you fill in? Take a guess: 100s, 1000000s, 1000…000s?
This problem enjoys many parable sources, e.g. a rich ruler promises a genius all the grains on the board then finds himself broke. The surprise growth shows up around the 10th square. Is there a formula to tell the total number of grains promised for the 10th square? Is there a way to calculate the number of grains in the last cell without adding all the grains in all the cells?
More generally, is “exponential growth” a productive way of predicting an outcome? We know about Moore’s Law and want to understand its implications as we review computing history. But the concept is all over the news as the Corona virus unfolds a modern parable of humanity, communities, travel industries, limited expertise, screwed-up governments, tests, and isolation to explore this topic.
Not everybody thinks in discrete math terms, but why not? Is there anything controversial about “exponential growth”? Is it deceptive? unbalanced? fruitful?
This is work I want to do gradually, maybe half an hour at a time. So, I’m publishing this article and will record future results separately. Catch up with this twURLed World if you’re interested or add your own trail exercises in the comments. See you later.
- Article: magazine 1945 “As We May Think”. See also “As We May Think” in Wikipedia. Recommended for its foresight, influence, and enduring value to red and re-read.
- About Vannevar Bush: Vannevar Bush from Wikipedia highlighted in Internet Archive 1995 Vannevar Bush symposium influences on mouse inventor (and more) Doug Englebart and hypertext dreamer Ted Nelson. Aside, whatever happened to the great domain name Vannevar.com and “Vannevar New Media”, an early web design company?
- Memex: Wikipedia Summary of Bush’s Memex concept and History of Memex;
60 years update on Memex, UColorado.
- Related Technologies:
Brown University Hypertext Symposium;
Semantic web (W3.org); Information Architecture Wikipedia;
issue-based information systems (IBIS) Wikipedia;
- Updates on Trails:
Jon Udell encourages annotation on Hypothe.is;
Clive Thompson wonders where trails start;
Steven Johnson links trails to writing and Dickens;
Weinberger question hyperlink virtue and more on Googlearchy.
- Exponential Growth Parable: Wheat and Chest Board.
Note: this bibliography shows several ways of referencing relevant material. Wikipedia generally offers good overviews. These references were selected to add depth and enduring value. The URLs are buggers to type and check using a screen reader. I’m working on a ‘link policy’.