A talk about a scientific paper is an invitation to read, not a substitute for reading
The talk
- Know the audience: adapt the talk to them
- Give the audience a reason to pay attention
- Do not present the content, but guide to understanding
- Plan the talk: tell a story
- Focus on the main message: omit unnecessary detail
- Keep it simple
- Clarity trumps generality
- Give examples
- Explaining too much is better than explaining too little
- Take about half the time to explain the context
- Be clear about what is yours and what is not
The slides
- Slides support the talk: they are not cue cards
- Every formula has to have an immediate purpose
- Allot time for each section of your talk
- Allocate two minutes per slide, on pain of rushing through the last few
- Number your slides, so that questions can refer to them
- Name your theorems, do not number them
- Text-heavy slides force the audience to read or to listen: they cannot do both
- Literature lists should be in the paper, not in the talk
Do not
- Do not go over time
- Do not be incomprehensible
- Do not overpack your slides
- Do not talk yourself down
- Do not talk other people down
The room
- Address the whole room. Tip: choose three people, one left, one right, one in the middle, and address them in turn
References