Reproducible documents and interactive graphics on the World Wide Web

Matthias Schwab

Author has no known email address

ABSTRACT

Reproducible documents, interactive graphics, and the Internet enables computational researchers to communicate and share their research easily. A reproducible document offers a reader three basic commands when inspecting any computational result: The first command, view, displays the result inviting the reader to scrutinize it. The second command, build recomputes the result from a set of given source files, such as experimental data and a processing program. The third command, info, offers the reader details on the result, such as the author's name, a time estimate how long a reproduction of the result may take, and possibly an URL which dispenses more information on the overall research project. My Internet version of a reproducible document includes the three reproducibility commands as pushbuttons. The underlying Java applet downloads and executes Java applications that process and display the results. A reader can step by step explore in depth a new technology before downloading and installing the technology's software. An author can maintain her collection of research results by routinely and automatically recomputing the results of all past research projects. A research group can verify the completeness and functionality of a software release of disparate research projects by recomputing all the results. For researchers programming in Java, a script can automatically derive Internet enabled reproducible documents from the standard products of a researcher: the written paper (e.g. in LaTeX), the programs (Java for Internet use), and a command script (e.g. a makefile or a shell script).