XML (2) Document Type Definitions (DTD)

XML提供了和显示无关的内容数据，可是数据的结构怎么确定呢？对于复杂问题，肯定是先有一个提纲，在给出具体内容，所以我们需要一种方式来提供XML的数据结构定义，这个就是DTD。

Element Declarations

<!  ELEMENT SAMPLE (#PCDATA) >

Document Type Declarations

<! DOCTYPE name_of_root_element SYSTEM "URL" [ internal DTD subset] >

external DTD subsets: Link to an outside file( often *.dtd)
internal DTD subsets: declaration in the XML file

Command-line validation
libxml2

Wed-based validation
www.cogsci.ed.ac.uk/~richard/xml-check.html
www.stg.brown.edu/service/xmlvalid

Elements Declarations

1. Define elements, like
   <! ELEMENT SAMPLE1 ANY >
   
   
2. Find out the #PCDATA, change all necessary elements to #PCDATA, like
   <! ELEMENT SAMPLE2 (#PCDATA) >
   
   
3. Find out the child elements, add to necessary elements, like
   <! ELEMENT SAMPLE3 (CHILD_SAMPLE1+, CHILD_SAMPLE2?, (CHILD_SAMPLE3* | CHILD_SAMPLE4| CHILD_SAMPLE5)?,(CHILD_SAMPLE6, CHILD_SAMPLE7)*, #PCDATA ) >
   

    Empty elements, like <!ELEMENT SAMPLE4 EMPTY>
    Comments in DTD, like <!-- comments -->

Tips:
Case sensitive
Sequence sensitive

Attribute Declarations
    <!ATTLIST Element_name Attribute_name Type Default_value>
    Attribute Types: CDATA, Emumerated, ID, IDREF, IDREFS, ENTITY, ENTINITYS, NMTOKEN, NMTOKENS, NOTION

<! ELEMENT AUTHOR EMPTY >
<! ATTLIST AUTHOR COMPANY CDATA #FIXED "TIC" >
<! ATTLIST EXTENSION CDATA #IMPLIED >
<! ATTLIST AUTHOR NAME CDATA #REQUIRED >

Namespaces
    Everyone can define elements, so name collision can be avoided. Like java, to avoid the class name collision, import Name Space.

All from 《XML 1.1 Bible 3rd Edition》