One way to look at elements and attributes suggests that elements are the 'real' containers of data, while attributes annotate elements with additional information describing the content of the element. The decision really comes down to a set of questions that are asked to determine whether we're dealing with an element or an attribute. Questions including:

• Will the item need to have sub-items? If so, only elements can have child elements nested within them while attributes can not.
  
  
• Will the item be a common data requirement of all users, that is, "core" information? If so, elements usually represent the natural, core content which would appear in every instance document while attributes represent data of secondary importance.
  
  
• Will the item describe data or the characteristics of data? Attributes typically are used to describe the characteristics, or properties, of an item.
  
  
• Does the data need to be ordered.
  
  
• What software will be processing and parsing the data? While element and attribute information is equally accessible in a tree-based interface, like those using the W3C's Document Object Model (DOM), developers working with stream-based interfaces, like the Simple API for XML (SAX) may prefer otherwise.

The committee recommends these guidelines when defining data items:

• An Element should be used to represent a piece of information that can be considered an independent item.
  
  
• An Element should be used when the information is related via a parent/child relationship to another piece of information. In this case, the element is also a sub-element of the element to which it is related.
  
  
• An Attribute should be used to represent any information "left over" after defining the items that have relationships to other items.
  
  
• An Attribute should be used to represent any information that describes other information, such as a status or ID.
  
  
• An Element must be used if an item occurs multiple times, because attributes can have only one occurrence in an element.

If it still not clear whether an identifier should be an element or an attribute, it is made an element. Elements are more extensible than attributes in an evolving standard because elements can contain other elements or substructures while attributes can not. If an item is defined as an attribute initially, and then needs to be expanded to hold fine-grained information, it must be changed to an element to be modeled correctly.