available to a substantial segment of the Earth sciences community. Submission, editing, and peer review will all be done electronically.
There is increased attention to electronic publication of astronomical research papers and data as well (see Box 3.1). Most conference proceedings are collated from electronic submissions in standard (e.g., TeX, LaTeX) formats. Abstracts of papers in most space science disciplines are now available on-line,24 and several journals are publishing electronic versions.
This trend is likely to accelerate and to open new opportunities for communicating research results to all scientists. Electronic publication will not just replace paper, however—it will alter the sociology of science. Writing, refereeing, and reviewing of a publication are now discrete and strictly ordered events, but they need not be in the electronic world. There, annotation, critique, elaboration, and revision can all go on iteratively and indefinitely, and in some instances no doubt they will. Some publications likely will become "living documents," under revision until they are no longer of interest. Even though our current social norms for attribution are based on the static publication model, it is doubtful that the scientific community would retain that model in order to preserve these norms. The value of a dynamic discourse is too great.25
Many electronic journals will not be "printable" in any meaningful sense. It is not just that they will contain motion and sound, but will incorporate also rich contextual links to the primary materials. Clicking on a graph will give the reader access to the data on which it is based, allowing alternative models and interpretations to be explored. A related important benefit of electronic publications is that results based on observations and modeling can be checked and validated by both reviewers and readers; restrictions on article length in paper journals and limited access to original data and software currently preclude any meaningful checks of the validity of published results based on observational data. A "copy" of the bits in an astronomical "plate" is as good as the original.
Publications also will become "active" agents, rather than passive stacks of paper. The term "program" has some of the wrong connotations, but nonetheless future publications will include executing programs-not ones that can be executed, but that are executing—autonomously gathering data, making predictions, becoming richer and more valuable as time passes.
In short, the Internet and World Wide Web are far more rapid and enabling means of communicating results, ideas, and other aspects of research than paper publications. Many changes in the conduct and dissemination of scientific research, from the individual to the international scale, may be expected to arise from these developments. 26
Related to the trend in electronic publishing is the increasing use of simulations and animations in research.27 Large-scale computation arose in part from a