Good talks some hints

1
Good talks some hints

• Henning Schulzrinne
• http//www.cs.columbia.edu/hgs/etc/talk-hints.htm
  l

2
Clear presentation goals

• Status update
• Milestones, deliverables,
• Discussion guidance
• IETF open issues
• Summarize somebody elses work
• candidacy, teaching
• Convince somebody
• (research) proposal
• Present own research
• conference, thesis proposal, project summary

3
Summary presentation

• Always answer
• what is the basic problem the paper is solving
• what is the core idea (algorithm, insight, )
• whats 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?

4
General hints

• Focus on concepts and motivation
• Dont 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

5
Content and presentation

• Add diagrams
• Network model, message flow, interactions
• RFCs often have too few!
• Dont just cut-and-paste the ASCII art
• Dont overload pictures use several similar
  ones
• Add tables
• Comparisons, summaries
• Use examples
• Allow listener to quickly grasp whats going on
• Confirms concepts

6
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

7
Dont

• Read slides aloud
• avoid full sentences or paragraphs
• Stand nailed to podium
• Ignore audience