Good talks – some hints Henning Schulzrinne http://www.cs.columbia.edu/~hgs/etc/talk-hints.html
Clear presentation goals Status update Milestones, deliverables, … Discussion guidance IETF open issues Summarize somebody else’s work candidacy, teaching Convince somebody (research) proposal Present own research conference, thesis proposal, project summary
Summary presentation Always answer Use common terminology what is the basic problem the paper is solving what is the core idea (algorithm, insight, …) what’s the approach (simulation, analysis, measurement, …) what is new about it Use common terminology may have changed since paper was published avoid introducing lots of symbols if used only once Use tables, trees, … to compare different papers Be careful with re-using drawings often too detailed for overviews Critical analysis any dubious assumptions - less general than claimed? conclusions sufficiently well-founded or stretching data?
General hints Focus on concepts and motivation Tie together concepts Don’t just repeat the paper or RFC Why is something being done that way? What value does this add? What assumptions are being made? Tie together concepts Why
Content and presentation Add diagrams Network model, message flow, interactions RFCs often have too few! Don’t just cut-and-paste the ASCII art… Don’t overload pictures – use several similar ones Add tables Comparisons, summaries Use examples Allow listener to quickly grasp what’s going on Confirms concepts
Slide mechanics No more than 6-7 bullets per slide Fonts smaller than 14 pt are unreadable Longer talks can have “bookmarks”, highlighting each section Gray-out the ones you already covered Use animation to focus on parts of a slide Avoid read-ahead by audience
Don’t Read slides aloud Stand nailed to podium Ignore audience avoid full sentences or paragraphs Stand nailed to podium Ignore audience