More thoughts on the seminar

Here's some more text I wrote in another email:
The goal is to have weekly informal talks about various general interest
computing topics (or practice talks for conferences). There are probably
three categories of talks that would be most relevant
I. The practice of research (including how to use LaTeX and Word
and the preparation of camera-ready versions)
II. HOWTOs on various research technologies (MPI, Berkeley Motes,
Globus, ... ). Here's where your talk would fit, and the goal
would be to cover some of the frequently encountered problems,
common solutions, and links to the best references on various
research systems that are commonly used by the different
research groups. Usually these will be systems developed outside
the department.
III. Personal research projects. For preparation of conference and
workshop talks. This provides a wider audience than each
students' research group so we encounter questions that don't
rely on the same assumptions we make in our own research groups.

The talks are meant to be short and informal. Slides
(powerpoint/pdf/keynote) are not necessary, but are suggested. If you
like, a short demo of the software would be fine.

15 minutes is probably the shortest amount of time that it would take to
get some useful information across, so if you need more time, then it
could probably go as long as 30 minutes.