From owner-InfoDesign@wins.uva.nl Mon Sep 13 12:00 PDT 1999 Date: Mon, 13 Sep 1999 11:52:38 +0200 (MET DST) From: multi-author To: InfoDesign and InfoGraphics Subject: InfoD: About Tufte's books Sender: owner-InfoDesign@wins.uva.nl Reply-To: InfoDesign@wins.uva.nl X-Organisation: - X-Address: - * Edward Tufte's books * (reactions) messages by: 1) Robin Kinross 2) Scott Confer 3) Doug Gillan _________________________________________________________________ 1) message by: Robin Kinross Roy Johnson asks for any thoughts from others about Edward Tufte's books, and refers us to his reviews at: http://www.mantex.co.uk/reviews/tufte-01.htm http://www.mantex.co.uk/reviews/tufte-02.htm http://www.mantex.co.uk/reviews/tufte-03.htm I find Roy Johnson's short reviews refreshingly plain-spoken, and they touch on worries that some of us have had about Edward Tufte's work on visual presentation. Tufte's discussion has splendid energy, has noble aims, has marvellous vigour in assaulting lies and deceit, and is full of unevenness and contradiction in argument - his absurdities, fresh insights, exaggerations, wisdom, are hard to disentangle. The books have the charm and the sense of isolation that so often marks the self-published production. (Here I speak from personal experience, as someone who has also resorted to publishing his own books.) Tufte's books have been so well and so generously produced - increasingly so, as they have followed each other - that this has seemed to disarm critics. Roy Johnson gets close to the problem, in the last sentence of his review of 'Visual Explanations': 'We might wish to query some of his theoretical claims, but it's very hard to be critically detached from such a seductive presentation of evidence - which paradoxically is the very point he's warning us about.' I tried to discuss all this at some length in a review-article about 'Envisioning Information' published in 'Information Design Journal', vol.6, no.3, 1991. After writing that review, I became aware of a few other such critical appreciations of Tufte's books, standing out from the uncritical gush that seems to constitute much of the literature on them, certainly from designers. I'd be interested to know of any up-to-date surveys of the discussion of Edward Tufte's work in information graphics. Robin Kinross _________________________________________________________________ 2) message by: Scott Confer Yes, Tufte's methods lack empirical basis. His data-to-ink ratio has been empirically studied by Gillan and Sorenson, and they found contradictory results: additional ink can enhance a graphic if done congruent with the meaning. This was one of several experiments conducted in the University of Idaho Human Factors Laboratory. The thesis by Sorenson was completed in 1993. The Human Factors and Ergonomics Society (HFES) has published guidelines for graph presentation: - Gillan, D.J., Wickens, C.D., Hollands, J.G. & Carswell, C.M. 1998. Guidelines for presenting quantitiative data in HFES publications. Human Factors. 40(1) pp 28-41 Scott Confer _________________________________________________________________ 3) message by: Doug Gillan I agree that Tufte has often received a lot of attention with little critical evaluation of his ideas -- evidence of the importance of packaging and marketing ideas. However, I also believe that Tufte has generated a number of interesting ideas that should be tested. There are some papers that have directly examined Tufte's ideas experimentally: - Carswell, C.M. (1992). Choosing specifiers: An evaluation of the basic tasks model of graphical perception. Human Factors, 34, 535 -554. - Casner, S.M. (1991). A task-analytic approach to the automated design of graphic presentations. ACM Transactions of Graphics, 10, 111 - 151. - Gillan, D.J. & Richman, E.H. 1994. Minimalism and the syntax of graphs. Human factors. 36(4). pp 619-644 I think that Tufte is chock full of ideas that might lead to interesting Master's theses. An example of this is Doug Sorensen's Master's thesis (University of Idaho). He looked at background ink in graphs (like background illustrations such as you might find in weekly newsmagazines) from a psychological point of view. He looked at the effect of these illustrations on performance of a graph reading task, on memory for information in a graph, and on preference for the graph. The typical experimental paradigm involved using variants of the illustration. For example, in the performance experiment, he had a no illustration baseline condition, an illustration that shared visual features with the type of graph (e.g,. box-like features if the graph was a bar graph), and an illustration that had different features from the graph (e.g., circular features if the graph was a bar graph). The data showed (although weakly as I recall) that the visually-similar background features made performance worse than the baseline, whereas the visually dissimilar background features improved performance. Doug found similar weak effects for memory (in that experiment, semantically similar backgrounds and graphs led to better memory than did semantically different backgrounds and graphs) and preference (a visually coherent background produced higher preference than the same elements arranged in a visually incoherent way). We never published these results. Doug Gillan _________________________________________________________________ This message was posted to both the InfoDesign mailing list and the InfoGraphics mailing list. _________________________________________________________________ From owner-chi-web@ACM.ORG Wed Sep 15 17:35 PDT 1999 MIME-Version: 1.0 Approved-By: instone@USABLEWEB.COM Date: Wed, 15 Sep 1999 17:18:00 -0400 Reply-To: thomas richardson Sender: "ACM SIGCHI WWW Human Factors (Open Discussion)" From: thomas richardson Subject: Re: Mappa.Mundi and Tufte To: CHI-WEB@ACM.ORG In-Reply-To: <37DFAA01.AE166683@crdp.umontreal.ca> Hmm. tufte never really says much about the Web in his books or in his lectures--and this is probably best. I think tufte is best used by Web info designers as source material rather than gospel. looking at his faculty bio (http://www.cs.yale.edu/people/faculty/tufte.html) tells you this guy is no Web guru--though his books are sublime. he may have a sense of where the Web should go, but how it gets there is up to the reader. the point about narrowly focused design is a good one. tufte produces books. books are a "medium of destiny"--they are fixed once printed (until my daughter rearrangesŠ but she stays away from tufte). I think of the Web more as dynamic than interactive, but you're right, tufte doesn't really do dynamic. that's okay. no one does dynamic yet. the value of tufte is that he pushes us to be mindful of the legacy of high resolution, densely packed information throughout the ages. his books are reminders that the presentation of information did not begin with the Web, and that we have a long way to go to even bring the Web up to information density standards of Gutenburg's time. -- The cover design diatribe is a bit awkward. Comparing tufte's covers to nielsen's is somewhat like comparing perrier to polywater. can we really expect "what sort of thumbnail is it going to make" to be the benchmark of cover art? for that matter, I don't feel that his covers display much information at all. seem pretty decorative, although the decoration is usually an illustration from the book. this example clouds your message. TR