Abstracts should be less than a page in any discipline (at least the ones I've encountered - computer science, maths, geography, linguistics, Eng. lit...). Goo

Abstracts should be less than a page in any discipline (at least the ones I've encountered - computer science, maths, geography, linguistics, Eng. lit...). Good "unstructured" (section-less) abstracts are about three paragraphs long and follow the structucture "what is the problem?" - "What is the state of the art regarding solution proposals?" - "Shortcomings of these that motivated this paper" - "What is proposed here as an alternative?" - [What are the key results?]. People are divided as to whether results should go into the abstract: some say the abstract should be a teaser that entices people to read the paper, others say the abstract should be a self-contained distillation of the paper itself. Structured abstracts are often nearly a page (common in the life sciences): they have explicit section headings like "Method", "Data", "Findings" etc., which is a great idea and should be embraced in all empirical disciplines at least, IMHO.