About the XML Files Found Here

The files in this directory are all examples under the didactic document type that is called article.

From early days of the project until mid 2006 the translator that converted the SGML form of an article to the (only) XML form of article automatically generated section numbers, tables of contents, and cross-referencing information.

Since mid 2006, this stage of the standard pipeline has been split in two, and there are now two XML versions of article. The one produced directly from SGML (with suffix “.xml”) is author-level and very nearly equivalent to original source. It resides under a strict document type definition for the XML version of article that may be regarded as suitable for authors who wish to write article originally as XML source. With the new standard pipeline these XML document instances qualify as “stand-alone” XML documents.

The second XML form of article (with suffix “.exml”) is very nearly equivalent to what was previously the only XML form of article in the production system. For maximum efficiency in subsequent generation of end formats the translator producing the second XML form simultaneously produces various auxiliary files containing (1) XML entity definitions (suffix .xet)1, (2) label information (suffix .xlb), and (3) a table of contents (suffix .xcn) as appropriate.

The items (1) and (3) represent included entities in the second XML version of an article and, therefore, must be present in order to produce conforming XML. This means that the second XML form of an article does not qualify as a “stand-alone” XML instance. The current formatters for the HTML and LaTeX targets make use of item (2).

In early 2002, motivated by CSS handling in the browser Opera 6, the author began very tentatively building a CSS style sheet for browser-based rendering of the XML version of the GELLMU article document type. The CSS sheet remains very much under construction, and, at best, CSS is a cruder rendering vehicle for arbitrary XML tag vocabularies than a finely honed translation to standard HTML or XHTML with appropriate use of CSS styling.

In general, there is an expectation that a web browser with XML capability will render an XML document that is accompanied by a CSS style sheet, but such browsers, as a general rule, should not be expected to handle document instances except when “stand-alone”.

In particular, the first XML form of an article is the one that should be considered most suitable for CSS-styled rendering.


  1. * Some of the entity definitions may be superfluous, but they cause no harm. (There is more than one conceivable way to generate the XML version of an article.)