This document contains a few suggestions for writing ICFP '96 submissions, all of which are intended to improve your paper's chances of acceptance for presentation at the conference.
The submission deadline is a firm deadline. All submissions received by the program chair after the date given in the call for papers (October 27, 1995) will be rejected without consideration. Please prepare your submission early enough to avoid problems with unforeseen delays.
ICFP '96 submissions must be full conference papers, but please do not format your submission in the two-column conference style required for final camera ready copies. Instead format your submission in single column format, 11 point on 16 point spacing. Do not attempt to bypass the page limit via the use of smaller margins, smaller typefaces, or less than optimal spacing. These techniques do not fool anyone and they make your submission more difficult to read. The following Latex 2e header should result in proper font size, margins, and line spacing.
\documentclass[11pt]{article}
\marginparwidth0pt\oddsidemargin0pt\evensidemargin0pt\marginparsep0pt
\topmargin0pt\advance\topmargin by-\headheight\advance\topmargin by-\headsep
\textwidth6.5in\textheight8.9in\renewcommand{\baselinestretch}{1.2}
Please include a list of one or two keywords or keyword phrases that will help the program chair assign readers for the paper. While a lack of such keywords will not directly affect a paper's chances of acceptance, it is always preferable for a paper to be read by people who are likely to understand it. If applicable, use keywords or keyword phrases extracted from the "suggested areas for submission" in the call for papers.
Although the submission need not be quite as polished as a final paper, please do not submit extended abstracts or refer to information "to appear in the full paper". Also, do not include appendices or other information that will not appear in the final version; the submission must stand on its own. Instead reference sources for more detailed information if appropriate.
If you are uncomfortable with writing in English, please have a fluent English speaker proofread the paper.
The submission should include an abstract that very briefly summarizes the contributions of the research described in the paper. This should be followed by an introduction that provides motivation for and a summary of the research and its results, targeted to someone who has not yet read the paper. Background information should be next, followed by one or more sections detailing the research and its results. Make use of figures and diagrams where appropriate to improve the presentation, along with descriptive captions. Related work should be discussed with background information or in a separate section near the end of the paper. The paper's conclusion should include a brief summary of the research and its results in specific terms for someone who has read the paper.
Do not make the reader guess the paper's contributions. Be precise and direct. Do not be modest, but do not exaggerate either. Admit any weaknesses. Write your submission in such a way that it will be accessible to and of interest to the entire functional programming community. The more generally applicable a result, the better.
Give an honest appraisal of its relationship to prior work. This is probably the single most important piece of advice in this document. Without such an appraisal, it is difficult for the reviewer to determine what is new and what is not. Do not omit comparisons with your own related work.
As stated in the call for papers, papers presented at the conference must describe new ideas or experimental results that have not previously been published. Furthermore, simultaneous submissions of substantially similar papers to more than one conference is unethical, and will be grounds for administrative rejection of a submission. Simultaneous submission to a journal is discouraged, but acceptable if the program chair is kept fully apprised of the progress of the journal submission. Presentation at small workshops whose proceedings are not widely disseminated is usually acceptable. If in doubt, contact the program chair. Papers presented at the conference are eligible for subsequent journal publication at the discretion of the journal editors.
Although written for authors of extended abstracts, many of the suggestions given in William Pugh's Advice to Authors of Extended Abstracts (SIGPLAN Notices 26, 6, June 1991, pp. 353-356; also available in postscript format.) apply equally well to full papers.
dyb@cs.indiana.edu