Results presented at conferences get immediate attention. Because of the slow turn-around time for journals and books and the fast pace of developments in CS, people working in the more active areas of CS rely heavily on conferences to get their work out. We distinguish between levels of conferences, based on the selectivity of the refereeing process.
Some conferences, especially those without multiple tracks, are very selective. It is not uncommon for such conferences to reject 75% or more of the papers submitted. Though journal turn-around times are longer, the final acceptance rates are higher than this. Therefore, acceptance of a paper at one of the more selective conferences may be counted as equal to or more of an accomplishment than publication of a paper in an archival journal.
The following criteria are useful in distinguishing conferences:
Conference papers are often later published, in revised and expanded form, as a journal paper. However, this is not always done. Sometimes the material is overtaken by newer results, and is absorbed into later papers, a book of collected papers, or even a textbook.
Conferences are important as a mechanism of technology transfer, i.e., publicizing and adapting research results to potential applications. In this regard, it is not uncommon for people doing CS research to submit papers to several conferences representing different communities of potentially interested users. This is viewed as a mark of successful research, that it has found multiple applications and potential avenues for further improvement.