Shelley’s posting Maids, Mommies, and Mistresses made me decide to throw in my own few cents on what makes a good conference submission, and how talks are accepted, to add to what Kathy Sierra and Adam Trachtenburg said (and there are good points in both). I’ve chaired a conference since 2001, organized tracks, and been a speaker at various conferences for many years, so I know something about the subject.
Number 1 has to be: if there are guidelines, read them and act on them! The conference organizers wrote them for a reason. I’m always amazed how many people obviously don’t read the ones I have for XML 2005 at Abstract Writing Hints — we get abstracts that are two sentences long, with misspellings, and acronyms used wrongly. The reviewers unceremeniously dump all of these.
I’ve been involved in lots of conferences and they range from the peer-reviewed to the “people we know or who pay get preference”; you need to figure out which conference you want to speak at and why, and which system they use, and how to have your talk accepted in that system. If the information isn’t on the conference web site about how talks are selected, email someone from the organizing committee and ask! Or find the name of a speaker from the previous year and ask them — most people don’t mind a brief polite email asking how they got on the program.
At the XML 2005 Conference I chair we use a blind peer review process to grade the abstracts. The Planning Committee then takes those grades and looks for program balance to cover interesting topics, knowing who the speakers are. This sort of system means that if you write a good abstract on an interesting topic, that isn’t topped by an even better abstract on a related topic, you’ll find yourself on the program. (Keynotes are a different story, of course, they’re invited). Most of the speakers each year are new speakers; some are “perennials” but that’s because they are involved in interesting work and know how to describe it in ways that make the reviewers want to attend the talk. The blind review system is biased towards submitters who can explain what they’re doing and why it’s interesting in 500 words or less, but I figure that’s a reasonable indicator for being a good speaker as well. It doesn’t always work that way (and we collect attendee reviews of the speakers each year to catch those cases), but usually it does. Oh, and another thing — it’s so much easier to have 100+ people help us figure out which talks are good than to rely on only 7 people on a Planning Committee!
The final piece of advice I’d give, once your talk is accepted, is to practise, if you’re not an experienced speaker. Even better, record your talk (audio and video) and watch the video to figure out what you can do better. Practise to yourself, the cat, or your family. Doing some professional training is good, but being familiar with the material so you’re not talking to the projected slides, or your notes, is better. Being prepared for likely questions is also good, and having a couple of “proposed” questions to give the chair of your session should nobody in the audience have questions never hurts. In other words, be prepared!