Friday, December 10, 2004

We're working on getting a globe display working in K-web (thanks to Jim Zaun, Gary Moyer and Bob Evans), but ultimatly we want to include the really stunning historical maps found at http://www.davidrumsey.com/view.html, which i note had much better WWWeb integration (ie the software you need is easy to use and loads itself). check it out!

Went to see Edward Tufte, the "da vinci of data" last night at Stanford.

A pretty good overview of ET work: http://www.edwardtufte.com/tufte/advocate_1099

Two fundamental issues: 1) tri-dimensionality and multivariables
2) information resolution (ie number of bits per time or area.)
The history of science is the increase in resolution (eg eye to telescope, now 10 to 44th or 100 millions times greater?)

Exemplar: Napoleon's march http://www.edwardtufte.com/tufte/posters
1. A wise answer to fundamental question "Compared to What?"
2. Shows causality (important in policy and intervention thinking; what allows us to survive in nature [and one suspects in "civilization"] is causal thinking)
3. brilliant because it shows 6 dimensions/variable (eg time and temperature)
4. multi-modal: picture, text, color, # etc (and all together, not separate as in standard report format with figs in appendix).
5. it documents everything, tells about it, and gives source of data
6. Cares about quality of content, as well as relevance and context (first maxim of info presentation is Do No Harm, which Powerpoint violates; see Shuttle example below, as well as http://www.dmkdmk.com/articles/tufte.html; classic PPT spook of Gettysburg: http://www.norvig.com/Gettysburg/making.html). ET notes that it took him 20 years to note that the name Napoleon occurs nowhere in the graph, b/c he was irrelevant.
ET claims good info design does not vary with time, gender, culture etc. A document should be designed to evoke an apporpriate kind of thinking, and shopuld not cater to fashion or even ease of use [//Engelbart?]

RELEVANCE TO KWEB: presumably his talk/work could be used in support of simple/minimalist design, but i see in it validation of giving kweb user quite a lot of data and flexibility in combining it (eg a timeline that can be expanded out to 6 kinds/strata of data, with images and color, which then interact with maps (eg where and when are these artifacts found?)
TE on interactive graphics: http://www.edwardtufte.com/bboard/q-and-a-fetch-msg?msg_id=00000s&topic_id=1&topic=Ask%20E%2eT%2e
an nice video spoof of glitz in presentation is viz-o-matic at the bottom of http://www.tc.cornell.edu/WhoWeAre/Outreach/index.asp

i have more extensive notes on the talk if anyone wants them