K' Scott Morrison

1
XML The Promise and the Reality

• K. Scott Morrison
• IBM Pacific Development Centre
• Vancouver

2
XML Hype
XML will replace HTML
XML is for documents
XML is for data
XML will replace all message formats
3
Profile Edit Screen NAME K. Scott
Morrison ADDRESS 8999 Nelson Way CITY Burnaby
STATE/PROV B.C. COUNTRY Canada
CODE V5A 4B5 TEL [phone]
FAX [phone] CREDIT CARD1 TYPE
VISA NUM 123456789 EXP 12/00 CREDIT
CARD2 TYPE AMEX NUM 987654321 EXP
04/01
4
Screen Scrape
Remote System
.. .. ..
Screen Scrape
Server
Persistent Store
5
Binary Representation
6
Binary Representation

• Issues
• Cant determine structure from data
• Portability
• Fixed field length
• Brittle interfaces
• Must modify all clients and servers
  simultaneously
• Mapping code typically buried in applications
• Significant maintenance problem
• Distribution of message map
• Not human readable

7
ANSI_X3.4-1968 (US-ASCII) Text Representation
8
US-ASCII Text Representation

• Standards-based, reasonably portable
• Human readable
• Can make conjectures about semantics
• Issues
• Limited character set
• NLS problems
• No real structure hierarchy, lists, etc
• Still very brittle
• Distribution of message maps

9
Proprietary Tagging
10
Proprietary Tagging

• Human understandable
• Less brittle interface
• Delimited text using tag and ltCRgt
• Issues
• Distribution of tag semantics
• Non-standard
• Character escaping issues
• E.g. ltCRgt, , etc
• No sense of hierarchy
• Programmer intensive parsers, handlers, etc

11
Formalized Tagging Markup

• Markup is meta-data
• Adds information about text
• What it means, how to interpret, how to render,
  etc
• Markup delimits
• START-----------END
• Interface is less brittle
• Markup works as a container
• Markup adds structure
• Hierarchy, etc

12
Markup Can Be Stylistic
Lorem ipsum dolor sit amet, consectetuer
adipiscing elit, sed diam nonummy nibh euismod
tincidunt ut laoreet dolore magna aliquam erat
volutpat. Ut wisi enim ad minim veniam, quis
nostrud exerci tation ullamcorper suscipit
lobortis nisl ut aliquip ex ea commodo consequat.
ltFONT FACETimes New Romangt Lorem ipsum dolor
sit amet, consectetuer adipiscing elit, sed diam
nonummy nibh euismod tincidunt ut laoreet dolore
magna aliquam erat volutpat. Ut wisi enim ad ltBgt
minim lt/Bgt veniam, quis nostrud exerci tation ltIgt
ullamcorper lt/Igt suscipit lobortis nisl ut
aliquip ex ea ltUgt commodo lt/Ugt consequat.
lt/FONTgt
13
Markup Can Be Structural
Lorem ipsum dolor sit amet, consectetuer
adipiscing elit, sed diam nonummy nibh euismod
tincidunt ut laoreet dolore magna aliquam erat
volutpat. Ut wisi enim ad minim veniam, quis
nostrud exerci tation ullamcorper suscipit
lobortis nisl ut aliquip ex ea commodo consequat.
ltPgt Lorem ipsum dolor sit amet, consectetuer
adipiscing elit, sed diam nonummy nibh euismod
tincidunt ut laoreet dolore magna aliquam erat
volutpat. lt/Pgt ltPgt Ut wisi enim ad minim
veniam, quis nostrud exerci tation ullamcorper
suscipit lobortis nisl ut aliquip ex ea commodo
consequat. lt/Pgt
14
Markup Can Be Semantic
Lorem ipsum dolor sit amet. consectetuer
adipiscing elit, sed diam nonummy nibh euismod
tincidunt ut laoreet dolore magna aliquam erat
volutpat. Ut wisi enim ad minim veniam, quis
nostrud exerci tation ullamcorper suscipit
lobortis nisl ut aliquip ex ea commodo consequat.
ltTITLEgt Lorem ipsum dolor sit amet. lt/TITLEgt ltBOD
Ygt consectetuer adipiscing elit, sed diam nonummy
nibh euismod tincidunt ut laoreet dolore magna
aliquam erat volutpat. Ut wisi enim ad minim
veniam, quis nostrud exerci tation ullamcorper
suscipit lobortis nisl ut aliquip ex ea commodo
consequat. lt/BODYgt
15
Simple Markup Example 1

• ltMessagegt
• Hello, World!
• lt/Messagegt

16
Simple Markup Example 2

• ltMessageContainergt
• ltMessagegt
• Hello, World!
• lt/Messagegt
• ltMessagegt
• Goodbye, World!
• lt/Messagegt
• lt/MessageContainergt

17
Profile Using Markup

• ltProfilegt
• ltNamegt K. Scott Morrison
  lt/Namegt
• ltAddressgt 8999 Nelson Way
  lt/Addressgt
• ltCitygt Burnaby
  lt/Citygt
• ltStateProvincegt BC
  lt/StateProvincegt
• ltCountrygt Canada
  lt/Countrygt
• ltZipPostalCodegt V5A 1B5
  lt/ZipPostalCodegt
• ltTelephonegt [phone]
  lt/Telephonegt
• ltFAXgt [phone]
  lt/FAXgt

18
Profile Using Markup (cont.)

• ltCardgt
• lt Type gt VISA lt/
  Type gt
• ltNumbergt 123456789 lt/Numbergt
• ltExpirygt 1200
  lt/Expirygt
• lt/Cardgt
• ltCardgt
• ltTypegt AMEX lt/ Type gt
• ltNumbergt 987654321 lt/Numbergt
• ltExpirygt 0401
  lt/Expirygt
• lt/Cardgt
• lt/Profilegt

19
eXtensible Markup Language

• Extensible
• Tag set is not fixed
• Structural
• Deep, hierarchical nesting of structures
• Ordered lists (unordered with Schema)
• Can infer meaning from structure
• Valid document requirement
• Can check structure against a schema
• Well formed (DTDs and Schema)

20
eXtensible Markup Language

• Portable
• Text-based Unicode
• All parsers must support UTF-8 (US-ASCII)
• Parsers may support UTF-16, EBCDIC, UCS-4,
  ASCII, ISO 646, ISO 8859, Shift-JIS, EUC, etc
• Human readable
• Machine understandable
• W3C standard
• Rich set of emerging tools

21
XML Tooling IE5
22
Isnt This Just Like HTML?

• HTML is a markup language based on SGML
• Key differences
• HTML has a fixed set of tags
• HTML mixes stylistic, structural, and semantic
  tags
• HTML does not support deep nesting and hierarchy
• HTML is invalid

23
XML Issues

• XML documents must be valid

Invalid

• Means that structure can be inferred

ltMessagegt Hello, World!
Valid
ltMessagegt Hello, World! lt/Messagegt
Hello, World! lt/Messagegt
ltMessagegt Hello, World! lt/MESSAGEgt
24
XML Issues

• Knowledge of document organization
• Distribution of document organization
• Character set issues
• Document focus
• Parser speed and complexity

25
Message Organization DTDs

• Document Type Definition
• Used to validate documents
• Issues
• Doesnt use XML syntax, not extensible
• Writing good DTDs is hard
• Very awkward and limited language constructs,
  uses eBNF grammar
• No namespaces
• No inheritance, defaults, ranges, enums

26
Profile DTD Example
lt?xml version"1.0" encoding"UTF-8" ?gt lt!DOCTYPE
Profile lt!ELEMENT Profile (Name, Address, City,
StateProvince, Country,
ZipPostalCode, Telephone, FAX, Card)gt lt!ELEMENT
Name (PCDATA)gt lt!ELEMENT Address (PCDATA)gt
lt!-- Lines removed for clarity --gt lt!ELEMENT
Card (Type, Number, Expiry)gt lt!ELEMENT Type
(PCDATA)gt lt!ELEMENT Number (PCDATA)gt lt!ELEMENT
Expiry (PCDATA)gt gt ltProfilegt ltNamegt K.
Scott Morrison lt/Namegt ltAddressgt 8999 Nelson
Way lt/Addressgt lt!-- Lines removed for
clarity --gt ltCardgt ltTypegt VISA lt/Typegt
ltNumbergt 123456789 lt/Numbergt ltExpirygt
1200 lt/Expirygt lt/Cardgt lt!-- Lines
removed for clarity --gt lt/Profilegt
27
Message Organization Schema

• Alternative to DTDs
• XML syntax
• Much more expressive
• Has namespace support
• Has inheritance, defaults, ranges (min/max),
  enumerations, sequences, unordered lists

28
Transforms XSL
ltProfilegt ltNamegt K. Scott Morrison
lt/Namegt ltAddressgt 8999 Nelson Way
lt/Addressgt ltCitygt Burnaby
lt/Citygt ltStateProvincegt BC
lt/StateProvincegt ltCountrygt Canada
lt/Countrygt ltZipPostalCodegt V5A 1B5
lt/ZipPostalCodegt ltTelephonegt
[phone] lt/Telephonegt ltFAXgt
[phone] lt/FAXgtltCardgt
ltNamegt VISA lt/Namegt
ltNumbergt 123456789 lt/Numbergt
ltExpirygt 1200 lt/Expirygt
lt/Cardgt ltCardgt ltNamegt AMEX
lt/Namegt ltNumbergt
987654321 lt/Numbergt ltExpirygt
0401 lt/Expirygt
lt/Cardgt lt/Profilegt
ltVisaCardgt ltCardNumbergt 123456789
lt/CardNumbergt ltExpirygt 1200
lt/Expirygt ltClientgt ltNamegt
K. Scott Morrison lt/Namegt
ltTelephonegt [phone] lt/Telephonegt lt/Clie
ntgt lt/VisaCardgt
Source Document
Destination Document
XSL Engine
29
eXtensible Stylesheet Language

• XSLXSL Transforms (XSLT) Formatting Objects
  and Properties
• Some basic XSLT functions
• Insertion of static text (like templates)
• Copy, discard, or rearrange source text
• Compute new text from source

30
XSLT Example XML to HTML

• lt?xml version"1.0"?gt
• ltxslstylesheet xmlnsxsl"http//www.w3.org/TR/WD
  -xsl"gt
• ltxsltemplate match"/"gt
• lthtmlgt
• ltbodygt
• ltPgt Address is
• ltxslvalue-of select"Profile/Address"/gt
  lt/Pgt
• ltPgt Name is
• ltxslvalue-of select"Profile/Name"/gt
  lt/Pgt
• lt/bodygt
• lt/htmlgt
• lt/xsltemplategt
• lt/xslstylesheetgt

31
XML as Data
ltProfilegt ltIDgt 123456789
lt/IDgt ltNamegt K. Scott Morrison
lt/Namegt ltAddressgt 8999 Nelson Way
lt/Addressgt ltCardgt ltNamegt VISA
lt/Namegt ltNumbergt 123456789
lt/Numbergt ltExpirygt 1200
lt/Expirygt lt/Cardgt lt/Profilegt
XML Document
XML-Db Extender
Profile Table
Database
Card Table
32
XML as Data Messaging
Request XML Message
ltgt ltgt lt/gtlt/gt
ltgt ltgt lt/gtlt/gt
RDBMS
Response XML Message
Server System
Client System

• Transport Examples
• HTTP over sockets
• MQSeries
• etc

• Message Infrastructure
• ebXML
• SOAP
• etc

• Message Formats
• OTA
• TravelFrame
• etc

33
Schema Distribution
Source System
Destination System
DTD
ltgt ltgt lt/gtlt/gt
Embedded
XML Message
ltgt ltgt lt/gtlt/gt
Replicated
DTD
ltgt ltgt lt/gtlt/gt
Centralized
Schema Repository
34
Applications Interoperability
35
Applications Interoperability
Thick PC Clients
MQSeries Integrator routing and transforming XML
messages
UNIX Servers
36
Applications eBusiness
Internet
Web/WML/XML Servers
37
Summary

• XML is here to stay
• There will be heavy vendor support for it
  because
• 1. XML is standardized
• 2. XML will be the B2B message format
• It can be leveraged now in most domains