The data objects also called as documents that conform to the syntax specification in XML are called Well-formed XML documents. These documents describe the structure, and are also known as standalone XML documents.

These documents are not dependent on external declarations, and attribute values receive no special processing or default values.

A well-formed XML document contains one or more elements in it that are delimited by the start and end tags. There is one element, the document element, which contains all the other elements within the document. All the elements are in the form of a hierarchical tree, thus, the relationship between the elements is in the form of a parent-child relationship.

So, to summarize, data objects are well-formed documents if :

Syntax conforms to the XML specification,
Elements are in the form of a simple hierarchical tree with a single node and
There are no external references to entities.

An XML parser that encounters a construct in XML and finds the construct not to be well-formed will report an error to the application as "fatal" error. This approach of error handling is the result of the compact design of XML and the intention that XML is to be used for much more than document display.

The W3C Recommendation has also described the behavior of Parsers or XML processor or the lower tier of the XML's architecture. This has been defined with the objective of easing the burden on the applications that handle the XML data.

There are two types of parsers namely non-validating and validating.

The non-validating type of parser merely ensures that a data object is a well-formed XML.

In the validating type, the parser uses a DTD to ensure the validity of a well-formed data object's form and content. Some parsers support both types along with configuration switches that determine the validation of the document.

The behavior of XML parsers has been defined with the purpose of easing the burden on the application's handling of the XML data. For example, the sequences of characters that are used as delimiters of the end of texts are operating system specific. Nevertheless, the XML application need not be concerned about this, as the parser will normalize all the delimiters to a single line-feed character. Whitespaces are another area where the parsers are constrained, as unlike HTML or SGML all whitespaces must be passed from the document to the application. The general entity strings are expanded by the parser as defined by the internal or external DTD subset.

XML parsers are required to normalize the attribute values (AttValue) before passing them to the XML application.

The table shows how the parsers handle the characters and references.

The AttValue is then processed by removing any leading or trailing spaces, and converting the multiple spaces into single spaces. The exception to this rule arises if the attribute value is declared as CDATA in the DTD and a validating parser is used.

There are two approaches in implementing an XML parser. They are the Event-Driven Parsers and the Tree-Based Parsers.

In this approach of XML processing namely the event-driven parser - the model which is familiar to the programmers of modern GUIs and operating systems - the parser executes a call-back to the application for each class of XML data that includes element with attributes, character data, processing instructions, notation, or comments.

Data handling in XML depends on the application as data is provided through the call-backs. The XML parser does not maintain the element tree structure, or any of the data after it has been parsed.

The most widely used structures in software engineering is the simple hierarchical tree.

In this approach, the well-formed documents are defined as a tree, and common and mature algorithms could be used to traverse the nodes of an XML document.

This approach conforms to the Document Object Model as specified by W3C. The DOM is a platform and language neutral interface that allows manipulation of tree-structured documents.

MSXML, a Java based XML, was developed by Microsoft. XML was later included as a part of the Internet Explorer 5 with a different parser.