Concepts, XML Document Structure

Tomáš Pitner, Luděk Bártek
<tomp@fi.muni.cz>
January 2015

Preamble

Lasaris

Fundamental requirement to all XML doc: it must be well-formed:

It contains prolog (heading) andexactly one root element.Hefore and after the root element, there can be processing instructions, comments (Misc).
It meets all the well-formedness constraints given in the specification.
Each of the parsed entities which is referenced directly or indirectly within the document is well-formed.

XML document structure we distinguish between: physical and logical structure.
Application programmers are usually interested just on the logical structure,
while for the authors of content, XML editors, processors may also the physical structure be important.

Logical structure: A document is divided into elements (one of them is the root), their attributes, text nodes in elements, processing instructions, notations, comments.
Physical structure: One logical doc may be stored in one or more entities; at least in the document entity.

are objects delimited by start- and end-tag, examples:

<body background="yellow">
   <hl>text node — content of element hl</hl>
   <p>text node — content of element p</p>
</body>

If an element is empty (no child elements, neither text content inside), then we write just empty element tag, eg.:

<tagname tagattributel tagattribute2. . . />

<hr width=' 507.' />

Or equivalently (from logical viewpoint):

<hr width=' 507.' ></hr>

"Attached to elements", carry "additional info" to elements - eg. its ID, required formatting (style) in case of (X)HTML, or links to other elements
Conceptually, we could replace all attributes with elements but we keep attributes to maintain readability.
The attribute content is NOT further structured (at least not according to XML standards. An application may see it other way but generally it is not recommended, cf. attributes in relational data model.)
The physical order of attributes in the start tag is NOT important and generally is NOT considered.

An attribute is composed of its name and value.
Attributes are inserted in the start tag which may be empty.
Attribute value is always in quotes or doublequotes add separated by a — from the attribute name.
For attribute names the same rules as for element names hold.
In one element, there can never be two or more attributes with same name.
If namespaces are used, neither two attributes belonging to the same namespace are allowed.

<table border='l'>
  <tr>
       <td>jedna</td>
       <td>dve</td>
  </tr>
  <tr>
       <td>tri</td>
       <td>ctyri</td>
  </tr>
</table>

They carry textual information, textual content.
Eg. in the next sample, the text ahoj ! is the text node - not the whole element em!
```
<em>ahoj!</em>
```

Similarly to HTML - comment is enclosed into <!—content—>
The comment content is content, NOT the the whole comment including markup.
Comments are usually not important for processing but it may depend on application, eg. Servlet-side Includes (SSI) use comments.
Parsers therefore should be able to forward comments to the applications.
SAX parsers ignore this in version 1!!! (resp. do so in version SAX2, in Java the package org. xml. sax. ext).

Entity is a basic unit of physical doc composition. Corresponds to string or file…
Parsers should process the entities such as the applications do not know about them.

We distinguish:

Document node: parent of the root element; may contain also Pis, notations, DOCTYPE etc. and
Root element: is the core part of an XML doc. In every file, there is just one.

in the next chapter XML family standards.