Analysis of the XML structure of Taylor & Francis journal articles should start with an overview of the main components of the document model.
The DTD recognises two kinds of article: "structured" and "unstructured". The latter kind is a "looser" model than the former, but contains all of the same components. It is intended for "pseudo" articles, such as book reviews (the rest of this overview focusses on the stricter model).
Article Segments
A structured article is broadly divided into three major segments:
However, none of these segments is held in a single container element, so the boundaries between them are not immediately clear. The following example shows the minimum content of a structured article:
<article>
<!-- FRONT MATTER -->
<meta>...</meta>
<title>...</title>
<shorttitle>...</shorttitle>
<abstract>...</abstract>
<!-- BODY -->
<intro>...</intro>
<part>...</part> <!-- or <section1> repeating -->
<part>...</part>
<!-- BACK MATTER -->
<references>...</references>
</article>
Frontmatter Segment
The article title is essential, as is a short version of the title which is used to ???. Also, every structured article must have an abstract. But the largest component is the article metadata ("metadata" is data about data, in this case information about the article, rather than the content of the article itself). The required components of metadata include the ISSN and Coden number of the journal the article is originally published in, the author of the article, the keywords that allow the article to be found by a precise search, and significant dates in the initial production of the article for primary publication:
<meta>
<issn>...</issn>
<coden>...</coden>
<author>...</author>
<search>...</search>
<production-dates>...</production-dates>
</meta>
Additional optional components include various additional titles (subtitle, super-title, journal title, abbreviation and DOI titles), a reference to an image of a logo, a set of keywords (identified by the author), an indicator of whether or not a table-of-contents should be produced, abbreviation and nomenclature sections (which can also appear in the backmatter), and an introduction to the article (though this is arguably the first part of the body content).
Body Segment
The body segment consists of one or more parts or sections. Parts are used when section numbering re-starts within an article:
<article>
...
<part>
<title>Part 1</title>
...
</part>
<part>
<title>Part 2</title>
...
</part>
...
</article>
or:
<article>
...
<section1>
<title>Section 1</title>
...
</section1>
<section1>
<title>Section 2</title>
...
</section1>
...
</article>
Backmatter Segment
The only compulsory component in the backmatter section is the block of references (though this could contain a single paragraph that explains that there are no references). However, this segment typically also contains summaries, glossaries and appendices:
<article>
...
<summary>...</summary>
<glossary>...</glossary>
<appendix>...</appendix>
<appendix>...</appendix>
<references>...</references>
</article>
Other components that may appear include acknowledgements, notes adding by the author during proof, abbreviation and nomenclature sections (which can also appear in the frontmatter), additional abstracts and translations of titles and abstracts.