The TFJA DTD was developed to guide and control the XML tagging of all journal articles published by Taylor & Francis. After initial publishing, the XML-based articles are stored in a company-wide archive for future re-purposing.
The DTD was initially created from the equivalent Dekker DTD (version 2.01), which was modified to remove support for document types other than Journal articles, and enhanced to support additional features found in other relevant DTDs (specifically, the Taylor & Francis US DTD and the NCBI DTD). The DTD is periodically updated to support new tagging requirements.
Note that there is no concept of a "journal DTD" within this DTD. Every article in a journal is considered to be a distinct entity that carries all of the information about the article, including details of the journal title and issue within which it was originally published. This approach allows articles to be separately flowed through production stages and individually stored in an archive of articles for re-purposing with other articles produced at other times, and from other sources.
Test Articles
Two example articles conforming to this DTD have been developed. The test articles were designed to serve two important roles:
| |
- |
to test all software that claims to be able to process or render articles that conform to the TFJA DTD
|
| |
- |
to serve as an appendix to this guide
|
In the first case, the articles contain at least one instance of every element declared in the DTD, including many potentially tricky combinations of these elements. Software that does not anticipate certain elements or combinations of elements will be properly "stress-tested" by these articles. This includes printing and online rendering products. Note that real-world examples should also be selected and used for testing software designed for this DTD.
In the second case, the article is largely "self-describing". This means that the text contained by an element generally describes that element (as in "<para>This is a paragraph.</para>"). Because the test articles contain at least one example of every element, in most or all of the contexts within which it can be used, a good understanding of the structure of articles can be gained by studying the structure and content of these articles.
These test articles are mostly the same, and vary to illustrate the different root elements:
The renderings may be viewed with full tags displayed, with brief tags displayed as hover text, or with no tags displayed. The XML file
999999999_suppl.xml is shared by both versions of the article.
Future Directions
The TFJA DTD will inevitably change as the articles are put to new purposes, and in response to issues that arise from its use in unusual circumstances. This section describes likely developments in the medium term.
Although a module containing MathML elements is included in the DTD package, none of the elements from this model are currently accessible from any of the elements in the main DTD. MathML is essentially "commented out". However, it is expected that MathML will be introduced at some point.
Although a module containing advanced chemical structure elements are included in the DTD package, none of the elements from this model are currently accessible from any of the elements in the main DTD. They are essentially "commented out". However, they may be introduced at some point.
It would be useful to be able to tag examination questions and answers in such a way that such exams could be extracted and used to help create an interactive learning experience. The exact tagging needed to achieve this is currently being investigated.