XML

Here are my notes on XML, the Extensible Markup Language.

Microsoft's FREE XML Notepad 2007 provides a simple intuitive user interface for browsing and editing XML documents.

Take the Brainbench certification test

Video XML Tutorials at Stylus Studio

Topics this page:
History/Summary
XML Protocol Stack
Infosetz
XSLT
XPointer, XLink
Your comments???

Related Topics:
Free Training! 
Technical Support 

Site Map  
About this site  

XML History

XML was first specified by the W3C “way back” in 1998 as a derivation of the more ambitious SGML (ISO 8879) serialization format that maps information to a byte stream.

XML is widely embraced because it can be read by every machine (and human). Plus, it's extensibile.

The content of XML models are defined by

• Document Type Definitions (DTDs), and
• XMI (XML Metadata Interchange) standard for expressing objects using XML. XMI specifies how to create XML schemas from models and reverse engineer models from XML documents, XMI 2.0 specifies how to create XML schemas from UML models.

The XML Protocol Stack

XML specifications are "layered" similar to the way the network protocol stack was designed in terms of isolated layers, listed here with the most abstract layer on top, progressing toward more concrete layers at the bottom.

Specification	#	XML Protocol Stack name	Enablers
Application specific	7.	Classes and objects
XML Schemas XSD definition language	6.	Types and Instances	XSLT defines transformations between schemas
XML Information Set (Infoset)	5.	Structural Items	SAX streams & DOM hierarchies project
XML 1.0 + Namespaces	4.	Elements and Attributes
XML 1.0	3.	Entities and Documents	XML Parsers
OS/Protocol specific	2.	Folders, Files, and Packets
Hardware specific	1.	Sectors and Bitstreams

Infosets

XML Information Set (Infoset)

Programming software referencing XML use one of two APIs which projectinfosets onto programmatic interfaces:

• DOM (Document Object Model) Node interface hierarchy allows dynamic "traversal" of the Tree Model

• SAX (Simple API for XML), a streaming (forward-only, read-only) interface for working with large infosets. It was cooperatively developed on the XML-DEV mailing list by a group of developers led by David Megginson

XML Formatting

All XML documents should begin with a prologue containing the XML declaration.

<?xml version='1.0' encoding='UTF-8'?>

XML is case-sensitive. Start and end tags must use the exact same case for the element name.

The most commonly used Unicode encoding in the English language world is UTF-8, an eight-bit Unicode transformation. Conveniently for English-speakers, the first 128 characters of UTF-8 correspond exactly to the ASCII character set. Another encoding, UCS-2, the canonical Unicode character set, uses the full 16-bit space for each character. Other commonly used encodings include ISO-8859-1 (which contains most Western European characters and is often called Latin-1), ISO-8859-5 (which includes Latin and Cyrillic characters), and EUC-JP and Shift_JIS (which encode Japanese.)

XML, like HTTP, is a markup language specification. But XML is designed to be extensible 3 allow developers to define new custom tags using phrases such as this:

<EMPTY/>
XML is stricter than HTML. XML tags must nest as subelements.

Companies & Products XML Diff and Merge Tool from IBM Alphaworks uses symbols and colors to identify changes to XML documents from a base document.

XML Parser for Java from IBM Alphaworks

Water, a new language for XML

RosettaNet

cXML, xCBL

XML Standards

Boston-based OASIS (Organization for the Advancement of Structured Information Standards) maintains the xml.coverpages.org and the United Nations CEFACT (Center for Trade Facilitation and Electronic Business), who are trying to standardize XML tag names — Business Transaction Protocols (BTP) by — for all business transactions being conducted on-line.
• The ebXML (electronic business XML) specification (draft v1.0.2 dated Jan 2001) is based on the Open-EDI Reference Model (ISO/IEC 14662). It has 3 parts:
  • Dynamically formulate trading partnerships through a registry and repository (RegRep or RR), which supplies a Trading Partner Profile (TPP) which define what ebXML processes are supported.
  • Negotiate and implement collaborative partner agreements (CPAs), formerly known as Trading Partner (TP) agreements. TPA/CPA data structures is adopted from IBM's tpaML.
  • Exchange electronic business transactions using a consistent XML-based messaging infrastructure -- formerly known as Transport, Routing, and Packaging (TRP)
• ebXML is in competition with XAML (Transaction Authority Markup Language) based on the eXtended Open Collaboration Protocol (XOCP) used by BEA's WebLogic Collaborate product.
• AuthXML and s2ml S2ML from Netegrity define the format of communications of authentication, authorization, and profile information. They allow interoperability of secure e-business transactions between disparate security systems.

• In March 2002, OASIS formed the XML Common Biometric Format (XCBF) Technical committee to develop a Common Biometric Exchange File Format (CBEFF) which consists of data element names necessary to uniquely describe parts of each human body —DNA, iris scans, hand geometry, fingerprints, face contours, etc. [ article]

  ebXML is being pushed by Oracle, Sun, and others in their camp.

Edifecs' CommerceDesk platform translates between flavors of XML.

David Brownell's (Sep 15 '99) "Conformance Testing for XML Processors" at XML.Com

$35 XML Handbook 4th Ed. (softcover with CD) (Prentice Hall, 2002) by Charles F. Goldfarb, Paul Prescod

$35 Oracle9i XML Handbook with "Oracle XML Developer Kit (XDK)" CD. (softcover with CD) (Osborne/McGraw-Hill, 2001) by Ben Chang, Mark Scardina, Stefan Kiritzov

XSLT

XSLT is a language for defining transformations between schemas.

XSLT relies on XPath (XML Path Language) for describing intra-document addressing.

XPointer (a language for describing inter-document addressing, points, and ranges) relies on XML Base.

Parsers

MSXML, Saxon, Instant Saxon, Xalan, Oracle XSL, Sablotron, XT, Unicorn, Napa, 4XSLT, and XML Pull Parser.

Assets

XPointer lets you specify subsections of documents

• XLink lets you connect documents and subsections

Google Web APIs Developer's Kit (dated August, 2002) provides a WSDLfile and examples in Java for using Google's SOAP interface and .NET for using Google's public Web Service.

Comment by Rael Dornfest Apr. 11, 2002

The XMLBench test driver for Java

SOAP Processing

A round-trip process:

1. serialize (marshall) data from objects into XML,
2. encode data for HTTP and SMTP transports,
3. bind to the remote object,
4. move data to the destination,
5. receive the response,
6. deserialize (unmarshalls) the response back into objects,
7. return the results to the calling method.

SAP SOAP Processor is used up to Web AS 620. From SAP NetWeaver 2004 on, please use the Web Service Framework.

SOAP Processing Performance

Frank Cohen published in Feb 2003 an IBM DeveloperWorks article presenting these results:

Among SOAP encoding styles adopted by technology providers.

• RPC SOAP Remote Procedure Call (RPC) encoding, also known as "Section 5 encoding" for where it's defined in the SOAP 1.1 specification (w3.org/TR/SOAP).

  SOAP RPC is the simplest for developers. With SOAP RPC, developers only need to code a single call to a generalized SOAP stack.

  However, the processing overhead of RPC-SOAP makes it the slowest of all.

  RPC is the default within IBM WebSphere Application Developer.

• RPCliteral (SOAP Remote Procedure Call Literal encoding) uses RPC methods to make the call as a single Tree but requires developers to marshall and unmarshall the body of the SOAP envelope.
• Doc-style (SOAP document-style encoding), also known as messagestyle or document-literal encoding, is efficient because an entire XML document is sent to a server and a return value is not required.

  This is the default within .NET development tools.

• StAX (Streaming API for XML)

The generalized SOAP stack provided by vendor are:

• IBM WebSphere 4 - Apache SOAP library.
  IBM WebSphere 5 - SAX-based Apache AXIS library.
• BEA WebLogic - BEA JAX-RPC APIs and one that implements the Java Web Services (JWS) APIs.
• SunONE Application Server - Java 1.4 JAX-RPC library

Other IBM DeveloperWorks articles on SOAP encoding:

Reap the benefits of document style Web services

Apache SOAP type mapping