Notes from Tufte
I went to one of Tufte's one-day courses. Apart from me being late, it was an interesting day. Most of the people in the (huge) room didn't seem all that interested. Judging by the general demeanour and dress, the course was heavily marketed at the corporate market. The good side of that was Tufte giving his 15 principles for giving good presentations.
Giving good presentations
You might not agree with all of these but they're certainly thought-provoking. One thing I noticed while listening to this is that, despite using very few visual aids (a couple of videos and his books), Tufte never gets boring, never stumbles, never ums or ahs. Quite an impressive thing to do for six hours.
Elsewhere during the day, Tufte complained about the Power-point style of presentation, referring to the idiocy of trying to convey information in bullet lists -- bullet lists don't allow you to "fill in between the bullets"; they show lack of thought, and lack of narrative. He also referred to the "strip-tease" style of presentation, and to "presentations that leave no trace." He was actually pretty funny at times.
- Show up early. "Something good is bound to happen." Watch your audience come in, introduce yourself.
- Never apologise. Stay out of the first person for a while.
- PGP -- particular, general, particular. Repetition helps to explain and helps retention.
- Give everybody at least one piece of paper (before the presentation starts!). It acts as a testimonial. It shows that you are willing to leave traces, to have consequences.
- Think of your audience in terms of what they read.
- [I lost this one: my notes say that paper gives a way to reason about information, to provide intensive content. He also said that paper was better than overhead slides.]
- Audiences are sacred. They are like colleagues; don't patronize them with Keep It Simple Stupid.
- Use humor. But use it to reinforce your points. Be careful not to gratuitously alienate your audience with your humor. But DO alienate them with your content!
- Don't use masculine pronouns as universal. Use the plural.
- Answer questions directly, don't embarrass the questioners. Demonstrate your mastery of detail. Side note: get the problem and solution you will present in the talk out at the start of the talk, so interrupters cannot destroy the talk. Count to ten when asking for questions. If concerned about lack of questions, plant them beforehand -- not overtly, but simply by chatting to members of the audience beforehand and mentioning certain topics -- in other words, open a channel of communication.
- The affect and impact of your presentation depends on the non-verbal element of how you communicate. If you believe in what you are saying, let the audience know.
- Finish early. "Something good is bound to happen." When was the last time you heard someone complain about a talk being too short??
- Practice, practice, practice. Record yourself. Watch the video only and look at your non-verbal style and mannerisms; listen to the audio only and listen to your use of "placeholders" (ums and ahs). Make sure you have the content nailed, so that you can reason about your presentation while giving it.
- Drink lots of water.
- And don't forget: most of the impact of your presentation depends on the quality, integrity, and relevance of your content.
Other notes
Not going to write down all my notes here, just a few key points.
- Information presentation should be effective but calm. Use the smallest effective difference.
- Always place information into context. Show the full data values. Use footnotes. Use annotation to bring causality to the data, using the minimal design necessary to link the words to the numbers.
- "Don't get it original, get it right."
- "Talent imitates, but genius steals."
- The BUGSAT method of interface design: bunch of guys sitting around a table. "We need a metaphor" is lousy design.
- "Design recapitulates bureaucracy" -- turf-staking shows up in the interface.
- Deep hierarchies are bad. Use a FLAT interface with lots of information content. Regarding web sites: people go there to find content, and there is an exponential drop-off in "drill-down."
- The "7 +/- 2" rule is nonsense. This was an experiment done by George Miller, and applies to the number of unrelated items that a person can keep in short-term memory. It was about memorizing, but we are interested in scanning. You need to show the overall scope to make scanning effective.
- Screen real-estate is computing's most precious resource. Don't waste it. We are reducing the resolution of the screen by design!
- Clutter and confusion are failures of design.
- Enforce visual comparisons; show causality; integrate text, annotation, graphics onto a single chart; presentation stands or falls on the integrity of the content.
- "Good design is clear thinking made visible."
- Beware of using anecdotes to support your case -- any anecdote can be "trumped" by another anecdote.
- Exhibits set the analytic agenda. Poor exhibits will impede/prevent clear thinking about the right topics. Think: what would I really like to see as evidence for that statement? How would I really like to see this displayed (out of thousands of possible ways)?
- When shown defensive presentations, ask "do you believe this or not?"
- Capture the imagination.
JohnR, January 26th, 1999