The DTD was developed from an earlier DTD (the Dekker DTD), and was hugely influenced by its design conventions.
Element and Attribute Names
All element and attribute names consists of lower-case letters.
When a name comprises two or more words, there is usually no separating symbol (for example, there is an element called "companyname"). However, there are exceptions to this general rule:
| |
- |
"name-text" = variants of highly structured elements that have a looser structure, allowing text to be inserted between child elements, and not forcing a particular order or occurrence requirements on those elements (for example "address-text")
|
| |
- |
"ref-name" = elements that hold citations of various kinds (for example, "book-ref")
|
| |
- |
"name-sect" = special category of document section (for example, "nomen-sect")
|
Block and Inline Elements
As with many DTDs of a similar nature, this DTD does not always clearly separate block-level elements from inline-level elements. It is possible to insert a block-level element within a block of text, thus splitting the outer block into two blocks with a third one between. For example, a paragraph can contain a list or table within it.
The logical explanation for this approach is that the text following an embedded list is not really a new paragraph, but a continuation of the one that started before the embedded list. Continuation paragraphs are often formatted differently, perhaps by starting at the left margin where the true paragraph starts with an indent. Another possible reason for this approach is to allow a complete logical paragraph to be easily isolated and extracted or presented as a complete unit of information.
All rendering software must be aware of this approach, and be tested to ensure that unusual text formatting effects are not produced in interrupted blocks.
Identifiers and Roles
Every element has two attributes that allow that element to be made the target of a link, and to allow it to be labelled for specifial purposes. These attributes are not expected to be used in production and publishing of an article, but may be exploited by document management and publishing systems.
While this concept must not to be confused with mechanisms that facilitate such functionality as hypertext linking or tying footnotes to their references, the same "Id" attribute is used for both features. In a few cases, the "Id" attribute is required for these other purposes.
External File Handling
This DTD allows for two very different approaches to referencing components of the article that are external to the main XML file, such as images. The "old SGML" approach is supported, where entity declarations are used at the start of the article to identify the location and data format of the file, and the "new HTML" approach is also supported, where a single attribute just names the file using a URL (the data format is indicated by the filename extent, such as ".gif"). Many of the systems that process articles must be above to work with both approaches and in some cases convert articles from one approach to the other.