The 
                  description element represents an alternate accessible description of an image, video, table, etc.
                  
               
               The 
                  description element can either contain the text content of the description or reference an external file in which it can be found via
                  the 
                  
                     xlink:href
                      attribute. It is not permitted to mix both methods, however.
                  
               
               To ensure the description can be rendered, the 
                  description element must not include descendant elements and attributes that require a description.
                  
               
               Usage details available in the Usage Details section.
               
             
            Allowed parents
            object
               
            
            Allowed children
            This element may contain text.
            
            This element may contain the following children: abbr, 
               address, 
               annoref, 
               annotation, 
               char, 
               citation, 
               code, 
               d, 
               definition, 
               emph (phrase variant), 
               emph (text variant), 
               expansion, 
               rend:linebreak, 
               ln, 
               name, 
               note, 
               noteref, 
               num, 
               object (phrase variant), 
               object (text variant), 
               quote, 
               ref, 
               s, 
               span (phrase variant), 
               span (text variant), 
               sub, 
               sup, 
               term, 
               time and 
               w
               
            
            Content model and additional requirements
            
               required attributes: 
@ref
                  
                
               optional attributes: 
@about, 
                  
@by, 
                  
@class, 
                  
@content, 
                  
@datatype, 
                  
@desc, 
                  
@its:dir, 
                  
@its:translate, 
                  
@property, 
                  
@rel, 
                  
@resource, 
                  
@rev, 
                  
@role, 
                  
@sel:selid, 
                  
@typeof, 
                  
                  
, 
                  
@xml:base, 
                  
@xml:id, 
                  
@xml:lang and 
                  
@xml:space 
               
               a choice of
                  
                  either no content
                  
                  or 
                     
                     
one or more of 
                        
                        a choice of either one of the 31 elements: 
abbr, 
address, 
annoref, 
annotation, 
char, 
citation, 
code, 
d, 
definition, 
emph, 
emph, 
expansion, 
ln, 
name, 
note, 
noteref, 
num, 
object, 
object, 
quote, 
ref, 
rend:linebreak, 
s, 
span, 
span, 
sub, 
sup, 
term, 
time, 
w or 
w 
                           or text
end of choice 
                        
                      
                     
                   
                  end of choice 
               Note that in addition to restrictions presented in the content model  above, 
                  					use of this element must also respect the following requirements:
                  
                     - The 
                        
description element must either contain text data or point to an external resource using the 
                        xlink:href attribute.
                        
                      
                     - The 
                        
description element must not contain text data and also point to an external resource using the 
                        xlink:href attribute.
                        
                      
                     - The 
                        
description element must not contain descendant 
                        description elements.
                        
                      
                     - The 
                        
description element must not contain descendant 
                        object elements.
                        
                      
                     - The 
                        
description element must not contain descendant 
                        table elements.
                        
                      
                     - The 
                        
description element must not contain descendant elements from the MathML feature grammar.
                        
                      
                     - Every 
                        
description element must be referenced by at least one element in the document.
                        
                      
                     - The 
                        
object element may contain only one child 
                        description.
                        
                      
                     - 
                        
description elements cannot be nested inside the children of an 
                        object element.
                        
                      
                     - A single 
                        
description element must be the only child of an 
                        object element when used.
                        
                      
                  
                  Such requirements take precedence over any conflicting statements in the
                     						content model or in the lists above of allowed children and parents.
                  
                
             
            Variants
            This element exists in 3 variants. The other 2  variants are:
            
            Namespace
            http://www.daisy.org/ns/z3998/authoring/
            Usage Details
            
            The 
               description element can be used in one of two ways: 
               
            
            
            
               
               - 
                  
                  
as a direct container for the descriptive text:
                  
                  
                     
<description by="republisher" xml:id="mouse-desc">
            <p>A picture of a small, grey mouse.</p>
</description>
                
                      
                  
                
               
               - 
                  
                  
as a pointer to an external description:
                  
                  
                     
<description by="republisher" xml:id="mouse-desc2" xlink:href="mouse-desc.xml" />
                
                      
                  
                
               
            
            
             An element may point to more than one 
               description so long as each description has a unique 
               role attribute value. 
               
            
            
             A single 
               description element can be the child of an 
               object to provide its accessible description. In this case, an implicit association is made between the two elements (i.e., the
               
               desc attribute does not need to be used: 
               
            
            
            
               
            <object src="mouse.png">
                <description by="republisher" xlink:href="mouse-desc.xml" />
            </object>
        
                
            
             If an 
               object element contains text or other elements as its children, an implicit 
               description is assumed. In the absence of the container element, the 
               object's children represent an implied content description added by an author. The following examples are functionally identical,
               but highlight the various implicit elements and attributes: 
               
            
            
            
               
            <object src="mouse.png">
                <p>A small mouse...</p>
            </object>
            
            <object src="mouse.png">
                <description>
                    <p>A small mouse...</p>
                </description>
            </object>
            
            <object src="mouse.png">
                <description by="author">
                    <p>A small mouse...</p>
                </description>
            </object>
            
            <object src="mouse.png">
                <description by="author" role="content">
                    <p>A small mouse...</p>
                </description>
            </object>
            
            <object src="mouse.png" desc="mouse-desc">
                <description xml:id="mouse-descrip" by="author" role="content">
                    <p>A small mouse...</p>
                </description>
            </object>
        
                
            
             The Content Selection feature can additionally be used to tailor the type of 
               description to the output format being generated: 
               
            
            
            
               
            <sel:select>
                <sel:when expr="target:format('BRAILLE')">
                    <description sel:selid="mouse-desc" by="republisher" xlink:href="mouse-braille-desc.xml" />
                </sel:when>
                <sel:otherwise>
                    <description sel:selid="mouse-desc" by="republisher" xlink:href="mouse-desc.xml" />
                </sel:otherwise>
            </sel:select>
        
                
            
             As each condition in the 
               sel:select contains the same 
               description element but corresponding to a different output, they cannot each have the same 
               xml:id. The 
               sel:selid attribute is instead attached to each 
               description. After the select condition has been evaluated, the remaining 
               sel:selid is converted back to an 
               xml:id.