The Operations Research Review Committee: David Burman, Hamilton Emmons, Arthur Geoffrion (Chair), John Muckstadt, Stephen Pollock, Richard Rosenthal
OR/MS Today, December 1991; reprinted in electronic form with permission
People won't read what they can' t read. In an increasingly technical field like MS/OR, the difference between merely serviceable writing and truly good writing often makes the difference between a manuscript that is accepted and one that is rejected, and between an article that is read and one that is not. Consequently, a high level of readability is required of all papers submitted for publication to ORSA and TIMS journals. Submissions displaying too low a level of readability are in danger of summary rejection.
Achieving a high level of readability does not mean avoiding formal rigor (in mathematics, logic or style), or avoiding necessary -- even complex or difficult -- technical detail. Rather, it means observing the basics of structure and presentation which one has usually learned in undergraduate English composition. Here are some specific ways to improve the readability of a manuscript being prepared for submission:
Introductory material should explicitly mention the operational phenomena of interest, even if the thrust of the manuscript is tactical or strategic. It should also motivate the reader to continue. It should state the nature of the problem studied. It should outline the manuscript's major substantive contributions and organization in a way that promotes accessibility to casual as well as serious readers.
The body of the manuscript should make generous use of metadiscourse -- discourse about what is being discussed at the moment -- so that the reader is kept well oriented concerning what is being discussed and why.
It follows that it is seldom sufficient to present a technical object (e.g., an equation, algorithm, theorem or proof) without an indication of how this object applies to the discourse at hand or why the means used to obtain it are pertinent. Indeed, if there is nothing novel or instructional, in, say, a proof or line of mathematical reasoning, then one should seriously consider excising it in favor of a brief remark or relevant citation.
Superior organization of the "storyline" and its supporting development is a sine qua non of readability. For example, anything in the body of a manuscript likely to divert attention from the current focus of the storyline should be omitted or put into appendices. Appendices are also often appropriate for the details of algorithms and proofs.
Technical terms and mathematical notation are, of course, frequently essential for precise communication or access to known results in science and mathematics. However, their excessive use will not fool referees into mistaking "rigor" for genuine substance, and can only diminish readability. Try to keep all notation concise and consistent. Some general rules are even applicable; for example, subscripts should not themselves have subscripts.
Figures and illustrations of all kinds are encouraged because visual thinking is powerful and efficient.
Graphs are usually superior to tables of numbers. Figure captions should be self-contained (identify all symbols, abbreviations and special terms). Ideally, an educated reader should be able to glance through a paper and discern its essence by reading the abstract, section headings, figures and typographically highlighted points (bullets, outlines, boldface phrases, summary tags, etc.).
Making a special effort to enhance a new manuscript's readability is a good investment of the author's time. It can shorten reviewing time, improve the odds of acceptance, enlarge the eventual readership and enhance the author's reputation. These suggestions, coupled with the "Instructions to Authors" provided by each journal, help point the way to increased readability. Progress in this direction benefits authors, readers and even our profession itself because MS/OR flourishes in proportion to how well we communicate our approaches and ideas to others.