I am writing a paper for a TCS conference and wondering what the community's opinion is on how abstracts in TCS should be written, good practices, or rules of thumb. I thought that I had read enough papers to be able to answer this question on my own, but after thinking about it a bit, I'm stumped.
Some specific questions I do not know the answer to:
- What is the purpose of the abstract? Is it to help those who do not want to read the paper figure this out as early as possible? To help a future researcher understand whether the paper contains a result of use to them? To communicate the key novelty/results/advancement of the paper? Something else?
- Who is the intended audience of the abstract? General computer science researchers, general TCS researchers, or researchers who specialize in the topic of the paper?
- Similarly, to what extent should technical terms be introduced/defined in the abstract?
- Does motivation belong in the abstract? Sometimes if I am not careful I find my abstracts turning into introductions.
- What is the value of listing precise results versus outlining the general question addressed? In other words, how specific should the abstract be regarding the results? It seems hard to hit multiple levels of specific/general without being redundant.
- When is it appropriate to mention/summarize prior work and background?
I was looking over responses to this question on examples of good TCS papers. There did not seem to be enough samples for me to learn, however :). The approaches to abstract-writing seem to vary widely.