TITLE: Describe the desired features of a "Mathematics Colloquium"?
QUESTION [14 upvotes]: I'm now a member of my department's colloquium committee.  Our task is to make a great colloquium series. I thought that the first step would be to come up with an appropriate definition of "Mathematics Colloquium." Standard dictionaries are not too much help:
Colloquium: n. 
 1. A usually academic meeting at which specialists deliver
    addresses on a topic or related topics and then answer questions relating to them.
 2. A conference at which scholars or other experts present papers on, analyze, 
    and discuss a specific topic.

So here's my question: What is the ideal definition of a mathematics colloquium?  Of course I know a bunch of things that make for bad colloquium talks, so I'm interested in knowing the definition of a good or great mathematics colloquium.

REPLY [4 votes]: Some suggestions:

Let all faculty members in turn choose speakers to invite.
This ensures that everybody gets involved, and that the interests
of all faculty members are represented.
Choose a room / building in which people feel well
(this is underestimated sometimes!), and choose a time where people
have time to attend and do not need to rush to their lectures
once the talk is over.
After the colloquium talks, always have a joint lunch / coffee / supper
(depending on the time of the day) in some pub / restaurant etc..
This fosters discussion among faculty members working on different
fields, and with the speakers. Try to choose a different location for this
every time.
As others have already said, tell speakers to give talks
which all faculty members can understand -- and not only
specialists in the respective field.