When we put human beings and computers together, we can do some wonderful things, but we can also get into some real messes. Usually, this is because we try to make information something it's not (information embellishment), or we destroy or incorrectly summarize data (information destruction). With some thought, we can avoid these two pitfalls. We will be looking at what can go wrong, and how we can do things right. We'll be drawing heavily on the work of Edward Tufte.

Information Embellishment & Clutter

Trying to make information more than it is.

Let's take a look at the Viz-O-Matic. It's not too much of a stretch - check this out...

Edward Tufte has coined a term for this - Chartjunk - extraneous material on charts and graphs which cloud their meaning. compare with Napoleon's March to Moscow.

A more subtle problem is computer "administrative junk" occluding information - see e.g. yearbook1.jpg (from Tufte's Visual Display of Quantitative Information)

Information Destruction

This involves either willfully or accidentally destroying data when trying to present it. It most often happens through incorrect summarization, particularly
  • "Dumbing down" information for executive use
  • Distorting information to fit into Powerpoint, Excel, or other computer tools
  • Sheer laziness or lack of thought

Here is one example - "Powerpoint does Rocket Science" - Conway's Law - instead use a report, like a commentary on a paper.

Another problem is that some very popular computer formats are "data destroyers" as they do not accommodate semantics, and thus information is lost.

"A common feature of all mainstream science publication is the universal destruction of high-quality information. Spectra, graphs, etc., are semantically rich but are either never published or must be reduced to an emasculated chunk of linear text to fit the paper model. The reader has to carry out "information archeology" using the few bricks that remain from the building. " (Reference)

See, for example, a Chemistry paper:

This has led to the idea of Datuments, Institutional Repositories and Nanopublications.

Getting it Right

So how do we do better? Well, we push for better interaction mechanisms for a start. Here is Edward Tufte's critique of the iPhone

Let's take a look at some other examples for more see Edward Tufte's Site:

Here are some tips
  • Read some of Edward Tufte's books.
  • Always look out for three pitfalls: information embellishment, clutter, and destruction
  • Consider presenting data and information for printing rather than screen (much higher resolution)
  • Consider using wikis or handouts instead of powerpoint.
  • Consider tables instead of graphs
  • Ask the question: will the user be thinking about the information or about the computer?
  • If you use data destruction formats like PDF, make the full data available in original form (or e.g. as XML)