After hearing all the group presentations this quarter, I have a few general
comments that almost everyone should keep in mind.

1.  Be interesting.  This is by far the most difficult goal to achieve but
    there are a few things you can do to help.

    a.  Be somewhat animated.  Change the pitch in your voice;
        challenge your audience; do whatever it takes to let them
        know something interesting is going on.

    b.  Try to gauge your audience's ability.  The best time to do
        this is at the beginning of the presentation.  Sometimes it
        helps if before you start your formal presentation you talk
        to the audience and get them interacting. This is where the
        common notion of making a joke comes from.  "Warm" the
        crowd up.  Get their attention at least for the beginning of
        your talk.  Ask them questions to find out who has heard about
        what you are talking about.  Based on their responses you
        can shorten or lengthen the background section.

    c.  If you do get the audience's attention initially, you have
        to try and keep it.  One way to do this is to make them
        interested in what you are talking about.  Give them some
        reason to want to listen.  "Ya know...  if you listen to 
        these safety steps, you could save your life during the
        next earthquake."  Okay, that's a bit extreme but I think
        you get the idea.  "Here's a protocol that could make your
        life a lot easier..."  This type of audience-tie-in will
        at least get you through the introduction.  Another way is
        to talk about what is going on in the industry.  "This
        protocol has just recently discussed in the IETF and they
        said <blah><blah> but before that it was an internal
        protocol used by IBM and ..."  A good story always works.

    d.  It is a well known fact that audiences love to listen to
        the intro, go to sleep for the main part of the talk, and
        then wake up for the conclusion.  If you have grand
        aspirations of keeping them awake during the main part of
        your presentation you have to work pretty hard.  One way
        that I have found to be somewhat successful is to give them
        a nice solid base that they understand and is easy to follow.
        This means a real-world example that everyone is familiar
        with and then tie whatever you are saying to that example.
        I mentioned this several times in the comments I gave to
        individual groups.  Explain how the protocol, framework,
        or whatever you are talking about is used for some common
        example.  Using an example also gives you a base case for
        comparing different techniques.  Advantages and disadvantages,
        especially as compared to some well known approach is a good
        way to give the audience a qualitative feel for what you
        are talking about.

     e. Sometimes, you do have to go into heavy detail.  But don't
        stay there long.  For example, audiences simply do not have
        the ability to understand a complex formula, no matter how
        slowly, clearly, and completely you present it.  There are
        two many variables and good formulas tend to be elegant and
        typically subtle.  Give the key details, tie them together
        with an example and then move on.  Make sure your transition
        out of the heavy details is very noticeable.  "Whew, that
        was complicated, but now I'm going to talk about..."


2.  The most important part of any talk is the organization.  You
    should spend more time organizing what you are going to say than
    any other part.  If you can find the perfect flow for your talk
    then the rest comes easy.  The most difficult part about
    organization is where to start.  There is always some body of
    background knowledge you need to present and there are some pretty
    tough constraints:  you can't present one thing until you present
    this other which requires knowing about this third thing, etc.
    Think about what you would get out of the talk assuming you knew
    only what the audience does and then ask questions about whether
    it makes sense to order your talk that particular way.

3.  And finally, be adaptive.  If you are presenting something and you
    see you have lost your audience then stop and get them back even
    if you have to sacrifice the details of your talk.  If they
    aren't paying attention they aren't going to get the details any
    way.  One lesson my advisor taught me about giving conference
    talks (which is obviously different than what we did) is that you
    cannot present the entire research in 15 minutes.  The only thing
    you can do is get them interested in it and get them to go off
    and read the paper.  The parallel here is you can't teach your
    audience everything about a topic but you can teach them that the 
    protocol, framework, system, etc. exists and communicate to them a 
    reason for investigating further and/or maybe even using it in the 
    future.