Fri Mar 28 13:58:56 EST 2003

My #1 piece of advice is to remember that the people reviewing the submissions are trying to read 100 papers in a week or two.  Therefore, rather than building up to your point (what I call "the surprise ending" method), start with your point (give away the surprise) and then use the rest of the paper to prove it, explain the terms and definitions, etc.

For example first paragraph would say, "We found that by using carbon instead of iron we've increased the tinsel strength of the alloy by 20%.  To do this, we had to overcome the heating issues normally associated with the process by using Superman's X-ray vision to control the experiment."  Sure, you've assumed that the reader knows what "tinsel strength" is, what "the heating issues normally associated with the process" refers to, and so on.  However, the people reading the papers DO know what these terms mean, and if they don't, they'd rather have them explained later in the paper now that they know to keep an eye out for definitions.  The people reading the papers, have been doing sysadmin stuff for a good long time so usually they know what you are talking about already.

This prevents the biggest problem that I've seen: we read through an entire paper and aren't sure what the person did.