Building a Better Table of Contents

As I near the completion of my thesis, I find it challenging to navigate through 250 pages of prose and remember where certain arguments or certain pieces of evidence are located; or challenging to get a broad overview of the narrative sequence.

Search is only effective if I can recall the name of reference or if the search term is suitably unique. Unfortunately technical terms are frequently repeated – for example – preattentive occurs 46 times. Even using keyword in context (KWIC) in the search panel still provides an excessively long list and perhaps the associated words of interest aren’t within a few words of the search term. Furthermore, search doesn’t work for the broad overview.

A Table of Contents should help, but not as much as I had hoped. A few problems:

  • On a large document, the table of contents is spread out across many pages, requiring paging and requiring the use of short term memory to hold onto relevant headings.
  • Headings tend to be terse, so it may not be obvious what the content is associated with a heading.
  • Furthermore, headings reduce information down to a few words – the narrative stringing the pieces of the argument together are lost.
  • While there are different levels of headings, overall, the representation of a table of contents is fairly uniform and undifferentiated, thereby not providing landmarks to facilitate navigation back and forth across the pages.

Table of Contents: an ordered list of headings without connective narrative and minimal landmarks spread across many pages. 

It may be suggested that the gist of document should instead be conveyed by the abstract. An abstract is more terse and provides the narrative sequence of the document – however – the abstract does not facilitate navigation of the document as it does not provide any links or page references. Furthermore, the length of abstracts tend to be quite limited: meaning that only the broadest overview can be described in the abstract.

Instead, I decided to use a very old technique. Chambers’ Cyclopaedia from 1728 creates a View of Knowledge – a unique narrative description of contents of his encyclopedia, all neatly organized in a hierarchical word-tree. Here’s a subset – you can read it left to right as a continuous narrative:


So, instead of a Table of Contents, perhaps Chambers’ approach is more useful. Here is a a narrative Description of Contents  for my thesis:


In one page, this description provides a brief readable narrative, outlining all the major arguments of the thesis, using FULL CAPS (and whitespace) to indicate major parts of the thesis. SmallCaps indicate each major heading. Superscripts (in blue) are links to page numbers.

[Aside: Note that word trees tend to use word size to indicate data. Unfortunately, word size uses alot of space; Full caps, small caps, superscripts and color can all be used to indicate data while offering higher data density and improved readability.]

The approach can be applied recursively: each part of the thesis has its own description of contents. For example, here’s the description of Part II:

In this example, the description contains many thumbnail images providing cues to the contents associated with each section (e.g. what’s the difference between text semantics and typographic semantics? – hint: comic book text is the image associated with typographic semantics). The large image at the bottom is a diagrammatic representation summarizing the design space derived in from the preceding sections.  Other cues (e.g. lines) can be added too to connect relations between items further apart.  Images also act as landmarks and commonality between different types of images help visually separate out various sections as well.

Does it work? I created the Descriptions to help me clarify how the pieces fit together. A couple of reviewers have commented how the descriptions have helped them understand the document – acting as a preview to each part and as a quick reference to navigate around the document. Obviously, more effort is required to create the description than a table of contents or abstract – but I’ve found it much more useful than either the abstract or table of contents to both navigate the document and to help move portions of the thesis into a better topics and flows.


About richardbrath

Richard is a long time visualization designer and researcher. Professionally, I am one of the partners of Uncharted Software Inc. I have recently completed a PhD in data visualization at LSBU. The opinions on this blog are related to my personal interests in data visualization, particularly around research interests related to my PhD work- this blog is about exploratory aspects of data visualization not proven principles.
This entry was posted in Data Visualization, Design Space, Search. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s