PublishingThe student is expected to publish in well-respected international peer-reviewed conference proceedings and journals.
A dissertation would most often be based on manuscripts corresponding to 3-5 publications in recognized international peer-reviewed outlets. Usually, results will be sent to conferences and journals in an on-going process during the Ph.D. studies. This would often mean that at the conclusion of the studies, a few publications are published (most likely in conference proceedings), and some are accepted or under review.
We want to emphasize that these are guidelines, since the publication process can be of a significant duration, and decisions should be made along the way by the students and the advisor. Too strict regulations would lead to suboptimal results. Below is an introduction to the publication traditions in Computer Science for people in other areas.
Computer Science Publication TraditionsTraditions in the field come from Mathematics as well as from the Engineering Sciences, so traditions vary, but in core areas, the following is typical.
Conference publication has significant merit. The meetings we attend, including the ones which are named ``workshop'', are typically conference style meetings. Generally, they do not have posters or one page abstracts. The meetings typically have the following properties: there is a program committee and a reviewing process, a significant number of submissions are rejected, accepted papers are included in proceedings (10-12 pages are typical) published by respected publishers or organizations, and the authors are allocated a time slot for presentation at the conference.
The turn-around time for a conference is about half a year from submission to publication (longer if the meeting uses post-conference proceedings). In addition, one may have to wait half to a full year for a (the) relevant conference to be announced. Since publication in the best conferences is prestigious, one often aims at a conference that is at least as good as the results merit. With acceptance rates between 10 and 50% for good conferences, this leads to rejection of sometimes excellent papers. In that case, one would often try a slighly less prestigious conference, adding up to another year to the publication time. After conference publication has been achieved, the paper is submitted to a journal in full (15-40 pages are typical). Turn-around time for journal publication is longer than for conference publication; a total of 2-3 years is typical.
Terminology from Other Areas