Step 1: Author writes a paper.
(See Halpern
and Trefethen's talk for details on this.)
Step 2: Author submits the paper.
There are two questions that will need to be decided at this point, namely
where to submit the paper and to whom. The former will be deferred until
later in the talk. The first thing to keep in mind when submitting the
paper is to follow the directions. In the back of most journals will be
instructions for submission. Do not violate these. Breaking these rules
runs the risk of incurring the wrath of the editor. There are generally
three possible options when deciding who to submit the paper to:
Step 3: The editor chooses reviewers/referees.
Reviewers and referees are anonymous to the author, thus it is
best, as alluded to previously, to submit papers to someone who knows
something about the subject. In most cases this will be the area
editor. Few authors do this, however, and often editors will receive
papers which are not central to their own area. The question now
becomes: How do editors choose reviewers for papers? Three strategies
are available:
Step 5: The editor sends reminders to reviewers.
After three months the editor will send reminders to reviewers. At some
point a reviewer may write back and say that they are too busy, or that
the paper is not in their area. If as a referee you are declining to
review a paper, it is considerate to offer some suggestions to the editor of
who might be a good referee. It's best that these people not be "famous,"
as the editor will undoubtedly have thought of the obvious choices,
but rather suggest people who are doing good work in the area, but are
not so well known. Note: the cycle of reminding reviewers and waiting may
occur more than once.
Step 6: The editor receives reviews and makes a decision.
When the reviews are returned the editor reads them and makes
a decision. Schneider notes that he never tries to determine if the
reviews are semantically correct. He prefers to operate under the
assumption that the referee is always right. This is actually fair because
if the referee is correct then the recommendation should
be followed. If the referee is making a mistake, for example saying
that the paper has flaws, then even if the paper doesn't have flaws it
needs to be rewritten and clarified so that the referee can be convinced
that it doesn't. In short, the onus is on the author to present the
material in such a way that it can be evaluated fairly.
When evaluating the reviews there are number of decisions about the fate of the paper that the editor can make:
Step 7: Author revises, review cycle begins again.
After the author revises the paper, assuming that choice was made, then
the editor will send the revised version back to the referees. In general,
the same referees should be chosen for each round of revisions. This is
more fair to the author, and also easier on the editor as well.
A typical paper will go through one major revision and one minor revision.
A minor revision may or may not involve the referees, as it may be possible
for the editor to determine that the appropriate changes were made.
The whole process from submission through reviews to being ready for
publication may take up to a year or 18 months.
Step 8: Acceptance and typesetting.
Once a paper has been accepted, then it goes into the queue of papers
that have been accepted to that particular journal. Once the paper is
sent to the typesetter, galley proofs may be sent to the author for a
final check. In general, an author has 48 hours to "okay" the galley
proofs. This is a real deadline, which, if missed, may cause the paper
to be dropped from the upcoming edition of the journal and pushed off
to the next one.
The author would like the submit-review-revise cycle to go as quickly as possible. And of course the referees would like their job to be as easy as possible. There are things that the author can do to facilitate this. First, write a clear and clean paper. Have other people read and critique it before submission. This is actually very important and could save at least one iteration in the review-revise cycle. After receiving comments from referees, it may be tempting to simply send back a revised version of the paper. A better approach would be to include a discussion of how you addressed the concerns of the referees. Make it very easy for the referees to quickly check that you made the changes that were asked of you. Of course, you will not always agree with everything that the referees propose. However, be sure to do something for every complaint if at all possible, even if they may be only cosmetic changes. Also, provide believable reasons why you're not making a certain change if you decide not to.
The question of where to submit your paper is obviously very important. Be familiar with the journals in your area. There are two classes of journals, those published by professional societies (for example, IEEE, ACM) and those published by companies. The former are probably better choices in most cases. Journals published by professional societies tend to be cheap to subscribe to, with a faster turnaround, a bigger audience and a higher standard. Moreover, editors for such journals are chosen by the members of the professional society. In contrast, journals published by private companies are more expensive and usually have a narrower focus. Also, the editorial board choices seem to be more politically driven than in professional societies. The thing to keep in mind is that in general a high quality editor will choose high quality referees who will probably give you more valuable feedback than "lower quality" referees. There are exceptions to every rule of course, and there are several good publications which are put out by private companies, and they shouldn't be overlooked.
Finally, Schneider offers some pointers on how to be a good referee. First, refereeing is anonymous, and referees should not talk about papers until after they're published. There are several facets to your job when writing a referee report. First, you need to advise the editor on what you think should be done with the paper. Second, you will need to establish your credibility both with the editor and the author. State what you think the main contribution of the paper is, explain what you think the problems are and justify your conclusions. Finally, give feedback to the author about the paper, with regard to both the exposition and the technical results. It is also perfectly fine to send a cover letter to the editor. Here you can say things to the editor that you may not want the author to see.
The review itself should be between one and five pages. It's clearly a problem if the review is longer than the paper itself. However, it's probably better to err on the side of being too detailed. Include a one paragraph summary of the paper. This establishes that you read the paper and what you understand it's contribution to be. Offer a succinct recommendation to the editor. Outline your major points of concern. The best form is probably to list them point by point rather than attempting to write flowing prose. Also number them, this will allow easy reference for you, the author and the editor later. You may want to follow this up with minor points (spelling, sentence structure, grammar and so on) or even mark a copy of the manuscript and send that back. (Although, sending a marked copy runs the risk of someone doing a handwriting analysis and figuring out who you are!)
When writing reviews, keep in mind that a distinctive style of writing or typesetting may not be such a great idea for anonymity purposes. Also, take care not to be too negative. Be professional and neutral in tone, not sarcastic. The cover letter to the editor is the place to vent, if necessary. Never say "you" when writing the review to refer to the author. Avoid such personalization and use phrases like "the paper," "the author," etc.
Ph.D holders will be both authors and reviewers, thus it is a good idea to have a solid understanding of the process, and what can be done both as an author and a review to make the journal submission process faster and easier on everyone.