Formatting Information

The reader will have deduced by now that while LATEX is possibly the best programmable typesetting system around, the LATEX file format is not generally usable with anything except the LATEX program. LATEX was originally written in the mid-1980s, about the same time as the Standard Generalized Markup Language (SGML), but the two projects were not connected. However, TEX and LATEX have proved such useful tools for formatting SGML and more recently XML that many users chose this route for their output, using conversions written in the languages already mentioned in § 8.2 above.

Unfortunately, when the rise of the Web in the early 1990s popularised SGML using the HTML, browser writers deliberately chose to encourage authors to ignore the rules of SGML. Robust auto-converted formatting therefore became almost impossible except via the browsers’ low-quality print routines.

It was not until 1995–7, when the XML was devised, that it again became possible to provide the structural and descriptive power of SGML but without the complex and rarely-used options which had made standard SGML so difficult to program for.

XML is now becoming the principal system of markup. Because it is based on the international standard (SGML), it is not proprietary, so it has been implemented on most platforms, and there is lots of free software supporting it as well as many commercial packages. Like SGML, it is actually a meta-language to let you define your own markup, so it is much more flexible than HTML. Implementations of the companion Extensible Stylesheet Language (XSL) provide one route to PDF via Formatting Objects with Extensible Stylesheet Language: Formatting Objects (XSLFO) but at the expense of reinventing many of the wheels which LATEX already possesses, so the sibling XSLT (Transformations) language can be used instead to translate to LATEX source code, as shown in the example in § 8.1.3 above. This is usually much faster than writing your own formatting from scratch, and it means that you can take full advantage of the packages and typographic sophistication of LATEX. A similar system is used for the Linux Documentation Project, which uses SGML transformed by the Document Style Semantics and Specification Language (DSSSL) to TEX

The source code of this book, available online at http://www.ctan.org/tex-archive/info/beginlatex/src/ includes XSLT which does exactly this.