Information Technology Reference
In-Depth Information
judge each point, because questioning as you write tends to stall the brainstorming
process. It can be helpful to set a time limit on this phase of no more than 20min.
In the second phase, assemble the talk by critically selecting the important points
and ordering them into sequence. During the second phase you should judge harshly
because otherwise the talk will contain too much material; be lean and leisurely, not
crowded and hasty.
A talk should be straightforward, although it can be used to convey complex
ideas. Rather than asking yourself what you want to tell the audience—the interesting
little issues explored, the technical problems confronted, the failings of the previous
research—consider what the audience needs to know to understand the main result.
Remember that a talk is a discussion with peers, not a news bulletin or political
speech. There should be a logical reason for the inclusion of each part of the talk.
Provide theminimumof detail that allows the audience to understand the result, while
being inclusive. If the audience believes that they have learnt enough to confidently
discuss the work with someone else, they will feel that the talk was of value. Think
of the talk as a demonstration that the work is of value and, in particular, that your
papers are worth reading.
Context can be as important as the ideas themselves. Take the time to explainwhy a
problem is important, where it arises, or why previous approaches are unsatisfactory.
Motivate the listeners so that they want to hear how a problem was solved.
Complex issues should be presented slowly and in stages; avoid detail that the
audience is unlikely to follow. Once listeners do not understand the flow of the
discussion, they are lost and will remain that way. Material that speakers shouldn't
present, although some choose to do so, includes messy details such as the internals of
a system, a proof of a theorem (attempting to walk the audience through a long series
of logical steps is a particularly bad idea), the elements of a complex architecture,
or technical information that is only of interest to a few specialists. There are of
course cases in which such material is necessary—the proof might be the main idea
to be conveyed, for example, or the theorem so unlikely that the proof, or its outline,
is required to convince likely skeptics—but as a rule the audience will not enjoy
being asked to understand intricate material that is unnecessary to appreciation of
the overall result.
Some material, particularly abstract theory, is dry and difficult to present in an
interesting way. However, there are alternatives to using the presentation to work
through the details of a theory. Rather than just discuss the research, explain the rela-
tionship of the results to the broader research area. Explain why the project was worth
investigating or consider the effect of the results on related work. Listeners who are
interested in the theory itself can speak to you privately or read the paper afterwards.
A talk is an opportunity to discuss problems. A speaker who is not frank about
shortcomings or difficulties, but is then exposed during questioning, can look foolish.
Obstacles are part of doing research and, not only can they add interest to a talk, but
just possibly the audience may offer solutions.
Never have too much material for the alloted time. Either you hurry through your
talk, not explaining the ideas well and getting flustered, or you run over time, the
audience is irritated, and the time for subsequent speakers is cut—not something for
which they will thank you.
