Apart from the fact that all XML documents that claim to conform to the TFJA DTD must be both well-formed XML documents and valid according to this DTD, Taylor & Francis have specified a set of additional rules on the application of XML markup.
XML Declaration
The XML declaration is required, and it must specify the "UTF-8" character set. The rules of XML state that, when present, this declaration must be at the very start of the file (the leading "<" must be the first character in the file):
<?xml version="1.0" encoding="utf-8"?>
Document Type Declaration
The DTD declaration must be present, and must include the system identifier:
<!DOCTYPE article SYSTEM "http://cats.informa.com/tfja/dtd/TFJA.dtd">
if the article root element is used, or if the unarticle root element is used:
<!DOCTYPE unarticle SYSTEM "http://cats.informa.com/tfja/dtd/TFJA.dtd">
The Taylor and Francis preferred public identifier may be included, but should be commented out and replaced with the system identifier when XML is delivered.
<!-- The Taylor and Francis preferred declaration -->
<!-- <!DOCTYPE article PUBLIC "-//Taylor and Francis//Journal Article//EN" "tfja.dtd" -->
<!DOCTYPE article SYSTEM "http://cats.informa.com/tfja/dtd/TFJA.dtd">
If entity definitions are included in the article, the end of the internal subset must be on a new line:
<!DOCTYPE article SYSTEM "http://cats.informa.com/tfja/dtd/TFJA.dtd" [
<!ENTITY .... >
<!ENTITY .... >
<!ENTITY .... >
]>
Complete Articles
Only entire articles are acceptable for new content, and complete article metadata headers with references are acceptable for digitization of back issues. One of the two article-level elements must be used:
<article>...</article>
<unarticle>...</unarticle>
The first of these elements should always be used for "proper" articles. Unstructured articles include book reviews, editorial, calendar of events, announcements and erratum.
Article Identification
The originator of the XML document is known as the "tagger" of that document, and must enter the name of the organisation (or an abbreviation of the name previously agreed with Taylor & Francis) in the "tagger" attribute of the root element. When an XML document is created, it should already be assigned a unique identifier, and this identifier must be entered in the "articleid" attribute.
<article tagger="Detica" articleid="123456789">...</article>
Line Feeds
Taylor & Francis request that hard returns be omitted from the content of tagged elements. Text inside of element tags should be allowed to wrap. In particular, hard returns should not appear in the text inside paragraphs. Please keep in mind that global deletion of hard returns after the fact can impact word spacing and will result in wholesale rejection of content delivery.
Hard returns can still be included between end and start tags of major components (such as paragraphs) for the purpose of readability of raw text legibility.