The basic language of scientific rules is coded to be platform independent. By knowing a few basic definitions, classes, and procedures, even aliens landing on earth should be able to quickly understand them. Consequently, a tool to communicate and simulate these rules should also obey this philosophy. Since it is unlikely the aliens could immediately run a Java-capable browser on their systems (though it would be one of their first efforts), I'll focus on human researchers around the world.

Pure research was one of the first areas to recognize the power of Java, as the Internet in its early stages was mainly a platform for scientists. The success of Java in science is only partly connected to the role of electronic publishing. Web pages became a new medium to present research results around 1992. The capabilities of the new medium are in a state of flux; but since 1996 with Java and contemporary browsers, these have reached a new stage. With enhanced versions, more and more scientific data is now being communicated through browsers (with or without plug-ins and add-ons) reading scientific files from public or private sites in standard HTML, Java, TeX, LaTeX, Postscript, DVI, PDF, and so forth.

Both writer and reader benefit from the new media. Preprints are sent to search-engines and servers (e.g., see arXiv.org), and people can read new results in their favorite browser shortly after the author has finished his or her last sentence. Soon, online archives combined with databases could become the primary form of reference (statistics from these servers show a strong increase in access rates).

Using the active links on a paper or the access to a searchable server is, obviously, much easier and faster than going to a library or to a copy machine. These active links are primarily used to connect readers to materials outside the publication and also allow them to jump inside the publication. Another advantage of electronic publications is the speed with which they can be widely disseminated. Part of the reduction in distribution time comes from shortening the refereeing process. To secure, verify, and document originality and accessibility, all files are mirrored and archived on servers distributed around the world. What is missing is the integration of a standard simulation platform into the TeX format. Implementing Java into scientific publications requires new standards. Tex or LaTeX could be another second-level Java platform. I suggest the inclusion of applet tags into the TeX language and to use a browser plug-in or a Java applet capable of displaying standard TeX/LaTeX scripts that include applets in a Java-enabled browser.