An OO schema language for XML

1
An OO schema language for XML
SOX
W3C Note 30 July 1999
2
What is SOX ?

• A schema language for defining
• the syntactic structure
• the partial semantics
• of XML document types
• Based on an Object-Oriented approach (Java)
• Developed by Commerce One

3
Why using SOX instead of DTD ?

• SOX is an efficient alternative to XML DTDs
• provide the same services plus
• SOX schemas are written in XML
• easier to read and to write
• can be managed using XML management tools
• an extensive and extensible set of datatypes
• embedded documentation
• extensive use of namespaces
• OO features reusability, polymorphism,
  inheritance

4
What is an XML schema ?

• A set of definitions providing a structured way
  of
• organizing data
• Defines structure rules using
• element type definitions
• datatype definitions
• namespace references
• .h file ? XML document .cpp file
• Characterized by an URI

5
Element type definitions

• Defines the basic structures found in an XML doc
• Close to classes in Java
• Tags in XML document instances
• Created using keyword elementtype

6
Element type definitions

• Characterized by
• A name unique identifier in a given schema
• A comment HTML text
• A collection of attributes
• A model
• defines structure and composition of an
• element of that type
• empty, string, element, sequence, choice

7
Element type definitions
ltelementtype nameparagraph"gt ltexplaingt
ltsynopsisgtThis defines the
paragraph elementlt/synopsisgt
lt/explaingt ltmodelgt
ltsequencegt ltelement
typestring"/gt
ltchoice occurs"2,"gt
ltelement typeint"/gt
ltelement typestring"/gt
lt/choicegt
lt/sequencegt lt/modelgt
lt/elementtypegt
8
Empty content model

• The element contains no data
• Useful when
• element is optional but its presence significant
• all the data is contained in the attributes
• ltelementtype name"NoContent"gt
• ltempty/gt
• lt/elementtypegt
• ltNoContentgtlt/NoContentgt

9
String content model

• Element contains only data, no other element
• Useful for simple structure containing only data
• String can be of any datatype
• ltelementtype name"StringContent"gt
• ltmodelgt
• ltstring datatype"string"/gt
• lt/modelgt
• lt/elementtypegt
• ltStringContentgtI love XML !lt/StringContentgt

10
Element content model

• Content restricted to be only of one type of
  element
• datatype or other element type
• Occurrence ?, , , (N,M)
• ltelementtype name"list"gt
• ltmodelgt
• ltelement type"item"
  occurs"2,9"/gt
• lt/modelgt
• lt/elementtypegt
• ltlistgtltitem/gtltitem/gtlt/listgt

11
Sequence content model

• Content expressed as a sequence of elements
• Element, sequence or choice
• Useful to ensure that data appear in a given
  order
• ltelementtype name"Sequence"gt
• ltmodelgt
• ltsequencegt
• ltelement type"boolean" name"Boolean"/gt
• ltelement type"float" name"Float"/gt
• ltelement type"date" name"Date"/gt
• lt/sequencegt
• lt/modelgt
• lt/elementtypegt

12
Choice model content

• lists a number of elements, sequences or choices
• Only one of the options appears in the instance
• ltelementtype name"Choice"gt
• ltmodelgt
• ltchoicegt
• ltelement type"string" name"String"/gt
• ltelement type"int" name"Integer"/gt
• lt/choicegt
• lt/modelgt
• lt/elementtypegt
• ltChoicegt ltStringgtThis is a string in a
  choicelt/Stringgt lt/Choicegt

13
Attributes

• Used to describe an element or its content
• has a is a relation to content
• ?
• nested element has a relation
• attributes of a class in Java

14
Attributes

• Characterized by
• A name unique identifier in a given schema
• A comment HTML text
• A type intrinsic or user-defined ? element type
• A presence specification element required,
• implied, default, fixed

15
Creating new attribute types

• Either derived from intrinsic type or from
  another
• user-defined type
• Created via keyword attdef
• 3 operators
• enumeration
• scalar
• varchar

16
Creating new attribute types
ltelementtype name"Car"gt ltempty/gt ltattdef
name"Color"gt ltenumeration datatype"NMTOKEN"gt
ltoptiongtRedlt/optiongt ltoptiongtBluelt/optiongt
lt/enumerationgt ltrequired/gt lt/attdefgt ltattdef
name"Registration"gt ltvarchar maxlength"7"
datatype"string"/gt ltimplied/gt lt/attdefgt ltattd
ef name"Mileagegt ltscalar datatype"double"
digits"6" decimals"1 minvalue"0"/gt ltrequired
/gt lt/attdefgt lt/elementtypegt ltCar color"Blue"
registration"MY CAR" mileage"72881.0"/gt
17
Namespaces

• Each schema exists in a namespace
• Can use definitions from other namespaces
• Import the external namespaces in the schema
• Libraries in Java
• Indicates original namespace using xmlns
• append prefix to the name of imported attribute
• and element types

18
Namespaces
ltnamespace prefix"containers" namespace "urnx-c
ommerceonedocumentContainer.sox1.0"/gt ltelem
ent name"Can" type"AluminumCan
prefix"containers"/gt ltelement type"PaperCup"
prefix"containers"/gt ltotherAluminumCan
xmlnsother "urnx-commerceonedocumentContainer
.sox1.0"gt ltotherPaperCup/gt
19
Inheritance

• Element type inherits structure from another
  type
• can add new structures using append
• can add new attributes
• Sequence, element and empty models only
• Can extend types from other namespaces
• Uses keyword extend

20
Inheritance
ltelementtype name"Ticket"gt ltmodelgt ltsequencegt
ltelement name"Price" typefloat"/gt ltelemen
t name"Time" type"datetime"/gt lt/sequencegt lt/m
odelgt lt/elementtypegt ltnamespace
prefix"tick"namespace"urnx-commerceonedocument
Ticket.sox1.0"/gt ltelementtype
name"MovieTicket"gt ltextends prefix"tick"
type"Ticket"gt ltappendgt ltelement
name"Title" type"string"/gt lt/appendgt ltattdef
name"PrePaid" datatype"boolean"gt ltrequired/gt
lt/attdefgt lt/extendsgt lt/elementtypegt
21
Polymorphism

• An element type specifies a specific type in its
  
• content
• An subtype of this type is used instead
• ltelementtype namePackage"gt
• ltmodelgt
• ltelement name"BaseTicket" typeTicket
  occ/gt
• lt/modelgt
• lt/elementtypegt
• ltPackagegtltBaseTicket PrePaidfalsegt
• ltPricegt15lt/Pricegt ltTimegt1123lt/Timegt
  ltTitlegtMatrixlt/Titlegt
• lt/BaseTicket gtlt/Packagegt

22
How does it work ?

• All SOX schemas conforms to the same DTD
• schema.dtd
• Everything is converted into Java Beans
• Reusing definitions is easy
• Enable robust distributed schema management