Introduction to RDL and REd

Research Description Language (RDL) is an XML application for describing and publishing scientific research efforts. Research Editor (REd) is a tool for editing RDL documents.

top

Rationale

One may wonder why anyone would need yet another tool for publishing scientific articles, since most scientists are perfectly happy using LaTeX. LaTeX is built on TeX, which is generally considered to be the best way to typeset complex mathematical formulas, but, especially in the form of LaTeX and other template packages, is now also being used for many other typesetting tasks. In fact, LaTeX has become the de facto standard for the communication and publication of scientific documents.

But LaTeX also poses its share of challenges to scientists:

By focussing on content, LaTeX claims to free its users from getting bogged down in visual details, as one would get when using a WYSIWYG text editor. Although it mostly succeeds in that goal, the price a LaTeX user must pay, is to get bogged down in command details. In fact, many people always have the LaTeX manual on their desk as they write their articles.
Next, visual editing does have some merits. By completely focussing on content, LaTeX is throwing the baby out with the bathwater. For example, the big, bold font used in section headings helps one navigating through a longer document.
Furthermore, the LaTeX commands are placed in the text and this distracts the user during his writing. One can therefore regularly see people write a bit of text, run LaTeX, and then check the output for errors in spelling, style, content, etc. This breaks the flow and reduces productivity.
Finally, and most importantly, LaTeX is a general typesetting program. It has facilities that are usefull for scientists, like figures and citations, but it isn't specifically targeted at scientists. For example, the only "fair" way of comparing different methods using multiple figures, is to have the same range and scaling on the axes of all the figures, and to have the figures near one another in the text. LaTeX offers no specific support for this task. In fact, to translate raw data from experiments to figures in LaTeX, one often needs to turn to other programs, like gnuplot.

Ideally, one would like a semi-visual editor that provides just enough visual feedback to be usefull, but does not fall into the WYSIWYG trap of editors like MicroSoft's Word. And this editor should have facilities to deal with scientific data from experiments in all sorts of ways. Oh, and it should typeset documents just as beautifully as TeX, and be able to publish documents in a variety of popular formats, like HTML, PDF, and PostScript.

This may seem like a lot to ask, and it is. But it can be done. The trick is to build an editor around a new format, based on XML. XML is a good candidate because it is both machine and human readable, contains both text and markup, garantees good handling of non-Latin character sets (since it supports Unicode), and because there are so many XML parsers available. We call this XML based format the Research Description Language.

top

Research Description Language

Research Description Language (RDL) is the next step up from Experiment Description Language (EDL). Whereas EDL is used to describe a single experiment, enabling others to replicate it, RDL focuses on combining results from multiple experiments in a coherent research effort. A typical RDL document contains references to multiple EDL documents, each describing a single experiment.

The strongest point of RDL is that it allows several standard operations on scientific data to be performed automatically by linking the data stored in EDL documents to tools that process that data. For instance, the edl2rdl-table program extracts the output part of an EDL document, and converts it into an RDL document containing a table. This program can be called automatically by adding the following code to the RDL document: <experiment src="example.edl" dest="table"/>. Analogous programs exist for converting data to figures, for calculating statistical properties like mean and standard deviation, and for determining whether differences are statistically significant. And still other programs can be added using the open architecture.

An RDL document looks a lot like a LaTeX document in at least one important way: it mixes text and commands for typesetting that text, which produces the same type of readability problems as discussed above for LaTeX. The true value of RDL therefore is not in the format itself, but in the fact that it can be easily parsed, displayed, and edited using a dedicated editor called the Research Editor.

top

Research Editor

The Research Editor (REd) lets the user edit his RDL documents semi-visually, and when the going gets tough, it translates the document to LaTeX sources for quality typesetting. Furthermore, it uses tools like pdfLaTeX, htlatex and divps for exporting to the respective formats. Support for other formats can be added using the open architecture.

REd's purpose is to let users edit RDL documents without them knowing anything about the XML elements and attributes that define RDL, or even about XML itself. It is a visual editor in the sense that the user need not type in commands but uses a graphical user interface with buttons and menus. It does not attempt to display the RDL document exactly as it will be typeset, however, i.e. it is not a WYSIWYG editor. But it does provide some valuable visual feedback where that helps the writer write his text.

top