XML Data and Process Modelling

1
XML Data and Process Modelling

• Fall 2005
• Thomas Hildebrandt
• Programming, Logic and Semantics Group, ITU

2
Road Map

• Why XML ?
• XML data model, XML Namespaces, XMLSchema
• XML Business Process ModelsWF and Com.
  patterns, the case of BPEL

3
Why XML?

• Models used for communication and
  interoperability by humans and programs
• inter-business and inter-application
• XML the de facto standard for tailor-made
  specifications of (semi) structured data
  collections shared and exchanged between humans
  and programs
• support for internationalisation (unicode) and
  platform independence (text based)

4
XML very briefly

• XML eXtensible Markup Language
• a merge of Hypertext languages and Markup
  Languages (SGML)
• text markup tagsltgreeting stylebiggthello
  worldlt/greetinggt
• a specific XML language constrains the use of tags

5
Recipe Example

• semi-structured mix of tags and unstructured
  text

6
Ordered trees
The underlying data model of XML is that of
ordered, node labelled trees
7
Node types

• text nodes
• element nodes
• attribute nodes
• comment nodes
• processing instruction nodes
• root nodes

8
XML recipe collection as tree
9
Textual representation

• properly nested, case sensitive tags
• attributes in-lined in start element
• predefined entities written as Unicode character
  references (e.g. lt as lt)
• comments as lt!-- .... --gt
• usually begins with an XML declarationlt?xml
  version1.0 encodingUTF-8?gt

10
Applications of XML

• data-oriented (e.g. replacing relational DB)
• document-oriented (e.g. XHTML)
• protocols and programming languages (e.g.
  XMLSchema, XSLT, WSDL, SOAP, BPEL4WS )
• hybrids (e.g. the RecipeML language)

11
XML Namespaces

• The XML namespaces mechanism allows one to
  combine XML languages
• ltexample xmlnsfoohttp//www.itu.dk/people/hilde
  /myfoospace xmlnshttp//www.itu.dk/people/hild
  e/mydefspacegtltfoofooelementgtlt/foofooelementgtlt
  /examplegt

12
Formal Language Specification

• needed for computers/programs to understand the
  language
• Consists of the Syntaxthe set of valid
  sentences and the Semanticsthe meaning of the
  valid sentences
• Often only syntax is specified formally...

13
Syntax

• Natural languages and programming languages are
  usually specified by a grammar (e.g. in
  Backus-Naur Form)
• Examples the 40-year old program Eliza, PASCAL,
  Algol 60 programming language
• Data collections are usually described by
  so-called schemas (e.g. databases and XML)

14
Schema languages

• A Schema/grammar can be described in a
  Schema/grammar language
• A schema processer/parser isused to validate
  (and normalize)a document

15
Why Schemas/grammars?

• Humans write documents that programs can parse
  and give meaningful error messages to invalid
  docs
• Why use standard schema language Standard
  parsers, program generation, ...
• Schema language should be expressive, efficient
  and comprehensive

16
XMLSchema

• Plenty of XML Schema languages, but DTDs and
  XMLSchema by far the most common
• XMLSchema is the most expressive of the two (and
  is itself an XML language)
• both uses regular expressions as subgrammar

17
Regular expressions

• Build from symbols, ?, sequencing, and
• Very effective (linear time, constant space)
• Very comprehensive
• Limited expressiveness Correspond to finite
  automata, that is, finite memory and thus can not
  express that a begin tag should be matched by an
  end tag

18
XMLSchema overview

• Namespace declarations
• Simple type definitions
• Complex type definitions
• Element declarations
• Attribute declarations

19
Student Records
ltschema xmlns"http//www.w3.org/2001/XMLSchema"
xmlnss"http//www.brics.dk/ixwt/students
" targetNamespace"http//www.brics.dk/ix
wt/students"gt ltelement name"student"
type"sStudentType"/gt ltattribute name"id"
type"string"/gt ltattribute name"score"
type"sScore"/gt ltsimpleType name"Score"gt
ltrestriction base"integer"gt ltminInclusive
value"0"/gt ltmaxInclusive value"100"/gt
lt/restrictiongt lt/simpleTypegt ltcomplextType
name"StudentType"gt ltattribute ref"sid"
use"required"/gt ltattribute ref"sscore"
use"required"/gt lt/complextTypegt lt/schemagt
Definitions and declarations populate the target
namespace
20
Simple type examples

• The student record example contains both
  primitive types and derived types ltattribute
  name"id" type"string"/gtltattribute name"score"
  type"sScore"/gtltsimpleType name"Score"gt
  ltrestriction base"integer"gt ltminInclusive
  value"0"/gt ltmaxInclusive value"100"/gt
  lt/restrictiongtlt/simpleTypegt

21
Derived types

• Patterns (regular expressions)ltsimpleType
  name"percentage"gt ltrestriction base"string"gt
  ltpattern value"(0-91-90-9100)"/gt
  lt/restrictiongtlt/simpleTypegt
• EnumerationltsimpleType name"passnonpass"gt
  ltrestriction base"string"gt ltenumeration
  value"passed"/gt ltenumeration value"not
  passed"/gt lt/restrictiongtlt/simpleTypegt
• UnionltsimpleType name"mark"gtltuniongtltsimpleTypegt
  ltrestriction base"passnonpass"/gtlt/simpleTypegtltsi
  mpleTypegtltrestriction base"percentage"/gtlt/simpleT
  ypegtlt/uniongtlt/simpleTypegt

22
XMLSchema declarations
lt?xml version"1.0"?gt ltstudent xmlns"http//www.b
rics.dk/ixwt/students"
xmlnsxsi"http//www.w3.org/2001/XMLSchema-instan
ce" xsischemaLocation"http//www.brics.
dk/ixwt/students
http//www.itu.dk/courses/IXMR/student.xsd"gt
id"hilde071072" score"97" lt/studentgt

• associates a Schema to an XML document
• may appear at sub-elements

23
Complex Types

• Complex types allow to specify elements and
  attributes as content of elementsExample with
  only attributesltelement name"student"
  type"sStudentType"/gtltcomplexType
  name"StudentType"gt ltattribute ref"sid"
  use"required"/gt ltattribute ref"sscore"
  use"required"/gtlt/complexTypegt
• elements as content may be specified by a regular
  expression? Element reference ltelement
  ref"t..." cardinality /gt? Concatenation
  ltsequencegt ... lt/sequencegt? Union ltchoicegt ...
  lt/choicegtand cardinalities minOccurs, maxOccurs
  "0","1","..", "unbounded"(default
  cardinalities are 1)

24
ltelement name"order" type"norder_type"/gt lteleme
nt name"email" type"nemail_type"/gt ltattribute
name"id" type"unsignedInt"/gt ltcomplexType
name"order_type" mixed"true"gt ltchoicegt
ltelement name"address"gt ltsequencegt
ltelement name"street" type"string"/gt
ltelement name"number" type"positiveInteger"/gt
ltelement name"floor" type"string"/gt
lt/sequencegt lt/elementgt ltsequencegt
ltelement ref"nemail"
minOccurs"0" maxOccurs"unbounded"
default"no email provided"/gt ltelement
name"phone" type"nphone_type"/gt
lt/sequencegt lt/choicegt ltattribute ref"nid"
use"required"/gt ltattribute name"priority"
default"normal"gt ltsimpleTypegt ltrestriction
base"string"gt ltenumeration value"normal"/gt
ltenumeration value"high"/gt lt/restrictiongt
lt/simpleTypegt lt/attributegt lt/complexTypegt
25
XML Summary

• de facto standard for tailor-made semi-structured
  data collections shared on the internet
• use namespaces to combine XML languages
• use XMLSchema to specify syntax
  formally,expressive - tends to be
  incomprehensible

26
Business Process Modeling(design and execution)

• Web services compositionacross organizational
  boundaries
• Technology push
• Need to improve efficiency
• BPEL4WS v1.1 Schemahttp//xml.coverpages.org/BPE
  L4WSv11-Schema-xsd.html

27
BPEL overview

• Process PartnerLinks variables flowchart of
  activities
• Primitive activities invoke, receive, reply,
  wait, assign, throw, terminate, empty
• Structured activities sequence, flow, while,
  switch, pick, scope

28
WF Patterns in BPEL

• Sequencing, Parallel Split Synchronisation
• Exclusive Choice and Simple Merge

ltsequencegt ltflowgt lt!-- activityA1 --gt
lt!-- activityA2 --gt lt/flowgt lt!-- activity
B --gt lt/sequencegt
ltsequencegt ltswitchgt ltcase conditionC1gt
lt!-- activityA1 --gt lt/casegt ltcase
conditionC2gt lt!-- activityA2 --gt
lt/casegt lt/switchgt lt!-- activity C
--gt lt/sequencegt (Example of condition
getVariableData(status)rejected )
29

• Multi-Choice and Synchronizing Merge(using the
  WSFL features)

30

• Cycles (using while loops)
• More detailed example (multible instances)

ltwhile conditionCgt ltsequencegt lt!--
activityA --gt lt!-- update condition data
--gt lt/sequencegt lt/whilegt
ltvariable namei typexsdinteger/gt ...
ltassigngtltcopygtltfrom expression0/gtltto
variablei/gtlt/copygt lt/assigngt ltwhile
conditionbpwsgetVariableData(i) !5gt
ltsequencegt ltinvoke .../gt ltassigngt
ltcopygtltfrom expressionbpwsgetVariableData(i)
1/gt ltto variablei/gt
lt/copygt lt/assigngt lt/sequencegt lt/whilegt
31

• Multiple Instances with Synchronisation

32

• Deferred Choice (using pick)
• Cancel Activity Using fault and compensation
  handlers
• Cancel Case ltterminate/gt action

ltpick namemessageortimeoutgt ltonMessage
partnerLink.. ...gt lt!-- activityA --gt
lt/onMessagegt ltonAlarm forPT10Dgt lt!--
activityB --gt lt/onAlarmgt lt/pickgt
33
Patterns not directly supported in BPEL

• WP8 Multi-merge
• WP17 Interleaved Parallel Routing
• WP18 Milestone

34
Communication Patterns

• Synchronous CommunicationRequest/reply and
  One-way (empty reply)
• Asynchronous CommunicationMessage passing no
  outputContainer
• No direct support for publish/subscribe nor
  broadcast

35
BPEL conclusions

• Anchored in the web services world
• Mix of flow/graph and control structures
• Expressive, perhaps too expressive
• Formalisation could provide precise semantics
• Not the only proposal (BPML, XPDL, WSCL,..

36
Course Evaluation

• Please spend the time... we do listen to your
  comments!!