A simplified subset of SGML (ISO 8879:1986) developed by a cross-industry group organized and led by Sun Microsystems
Not new: builds on 30 years of research and 14 years of SGML standardization
Offers powerful data modeling -- no limits on namespace or structural depth
But small enough for Web browsers
Not a language but a metalanguage
Designed to support the definition of an unlimited number of languages for specific industries and applications
All XML languages can be processed by a single lightweight parser
Tags have no predefined meaning
Unlike HTML, XML by itself conveys only content and structure, not presentation, behavior, or meaning
The meaning of XML languages must be specified outside of XML itself
Operational semantics: programs, servlets, applets, scripts, stylesheets...
Definitional semantics: prose, namespaces, ontologies, UML diagrams...
Legal XML documents are called
A well-formed document describes a
If a well-formed document conforms to an optional grammar or
schema (e.g., a DTD), it is also
Well-formedness is a property of the document
Validity is a property of the relationship between a document and a grammar or schema
XML provides a standard framework for making agreements about communication.
Industry DTDs and schemas
Industry namespaces
But XML does not make those agreements. That would be magic!
Agreements are made by people working together on a shared problem. This is the process of language standardization.
XML has a social agenda: freedom from vendor control.
(1996) We knew that HTML could not support data exchange in general
Without XML, HTML would have been replaced with a binary, proprietary format controlled by a single vendor
The alternative was SGML (international, open, text-based standard, ISO 8879-1986)
XML put SGML on the Web. Result:
Users can define their own data exchange standards
There are many inexpensive, robust tools
In theory, data belongs to the people who create it rather than to the software vendors
In practice, freedom depends on the open standardization of XML tag languages
XML standardizes the
Deploying XML creates an
Standardized parsers for putting data into memory
Standardized interfaces (DOM and SAX) for processing the data
Standardized ways to display data (CSS, XSL-FO)
Standardized ways to query data (XPath, XQuery)
Standardized ways to link data (XLink, XPointer)
Standardized training of people in both publishing and data processing
A document is data that you can
Document requirements are a
The basic problem with documents is that we need to display them
in many
Example: This presentation.
Written in XML
HTML generated using a stylesheet for online publishing
RTF generated using a different stylesheet
The SGML/XML publishing model decouples data from processing
This isolates changes in large systems, making them more flexible and reliable
Basing a system on XML makes it well-suited to transactional processing in a heterogenous, asynchronous, distributed environment (like the Web)
XML uses documents as the transfer mechanism for data. XML is text.
Creates large, standardized collections of data that can be processed by simple text tools
Creates an infrastructure in which human beings are still part of the process
Troubleshooting
Generation of human-readable deliverables (e.g., catalogues)
Integration into existing social institutions
Promotes a message-oriented view of electronic commerce that isolates business transactions from differences in software, hardware, system architectures, and application programming languages
XML was designed to enable many different views to be generated from the same set of data using stylesheets.
So it is perfectly suited to
Report generation
Display of business data to consumers
Display of business data to programmers and business professionals
Various forms of electronic delivery (e.g., Web sites)
But the markup language must be designed to make this easy.
Joint 18-month effort of OASIS (the Organization for the Advancement of Structured Information Standards) and UN/CEFACT (United Nations Centre for Trade Facilitation and Electronic Business)
Over 1000 participants from all over the world
The vision: A global electronic market place where enterprises of any size, anywhere can:
Find each other electronically
Conduct business through the exchange of XML based messages
Standard messaging services for XML
Standard registry/repository for storing business models and profiles
Standard mechanism for trading partner agreements
Standard architecture for forming trading partner agreements and exchanging business messages
We believe in an open approach to electronic commerce
We believe that no single company or country should dominate the infrastructure for electronic commerce
We believe in empowering users
ebXML is a nonprofit, independent effort sponsored by the United Nations
ebXML is the only existing non-proprietary framework for electronic commerce
I am proud to say that Sun has been the single strongest industry supporter of the ebXML effort
Everyone who believes in open standards for electronic commerce should support ebXML.
I believe that the evolution of XML-based electronic commerce will take place in three phases or levels of implementation.
Standardize the infrastructure (messaging, discovery, trading partner agreements)
Standardize the messages (business documents: purchase orders, invoices, ...)
Standardize machine-processable semantics (business process models)
With the completion of the ebXML infrastructure specifications in May 2001, messaging, discovery, and trading partner agreements are now standardized and can be automated.
Most primitive implementation: email ad-hoc XML, conducting all other aspects manually using email or phone calls
This would usually be a one-sided relationship
But it allows the smallest businesses to participate
In Level 1, the role of the repository is to supply basic infrastructure specifications and commonly used industry tag sets
Level 2 adds standardized XML messages defined by humans within some standardized semantic framework using standardized core components (data elements)
Enables off-the-shelf software solutions
Enables a transition from existing business practices by replacing traditional business documents with human-readable XML equivalents
Many options for defining the semantic framework
UML
Prose
Prose + DTD
In Level 2, the repository will include descriptions of standard business models, standard core components, and standard XML business documents
In Level 3, standardized messages are automatically generated from a standardized semantic framework
Modeling done from the top down
Schemas generated automatically
The whole system operates without human intervention
The repository now includes business models designed for automation
Advantages:
Completely automated
Allows dynamic modification of business models
Problems to be solved:
Development of the technology itself
Entry point for SMEs
Transition from existing business and legal systems
Standardize the infrastructure (messaging, discovery, trading partner agreements)
The ebXML infrastructure provides this level of functionality
Standardize the messages (business documents: purchase orders, invoices, ...)
This is the next stage of development
Standardize machine-processable semantics (business process models)
This is a long-term project for the future
In level 2, standardized business documents take their meaning from standardized business models. This diagram shows the role of planning and shipping documents in a business model from the automotive industry.
To achieve the next level of e-commerce implementation, we need to standardize a common business language.Examples of what needs to be standardized:
Date
Weight
Price
Address
Telephone number
Party
Purchase order
Shipping notice
Invoice
Atomic components
Date
Weight
Price
Aggregate components
Address
Telephone number
Core blocks
Party
Documents
Purchase order
Shipping notice
Invoice
"Standard" data components change structure when put into different contexts.
Business process
Industry
Product type
Geographical region
Regulatory environment
Role (vendor, customer, etc.)
Even such a simple data item as an address will change depending on the context.
Addresses in Japan are different from addresses in the United States
Addresses in the auto industry are different from addresses in other industries
Users are rushing to adopt various existing XML business syntaxes.
cXML
xCBL
RosettaNet
OAG
BizTalk
various other dialects
The existence of multiple business languages is a serious interoperability problem. But once adopted, these existing vocabularies will be very hard to dislodge.
Put one of the existing XML b2b libraries into a genuine standards process.
Populate the process with XML and business domain experts.
Persuade other groups working to standardize XML business grammars to join the common effort.
Modify the default library as necessary to achieve consensus.
Lead users into adopting the standard "Universal Business Language" (UBL)
Evolve the standard UBL as we gain experience.
Optimized for Internet b2b
International in scope
Applicable across any sector or domain of electronic trade, transport, and administration (purchasing, payments, logistics, transportation, statistical reporting, social administration, healthcare, etc.)
Interoperable with existing EDI systems
Based on a core library
Unencumbered by intellectual property claims
Already proven to work in actual electronic marketplaces
Proven implementability
Easy testability
xCBL 3.0 represents four years of work that we don't want to start over from scratch.
xCBL's IP is unencumbered. This is really important.
xCBL 3.0 reflects experience gained in EDI, RosettaNet, OBI, and ebXML.
Microsoft and SAP are basing systems on xCBL.
Many vendors already have xCBL code for use in adapters.
The owners of xCBL are willing to put it into a real standards process for the benefit of the industry and to work within that process to advance the industry standard.
In the software business, this is as good as it gets.
The basic idea is to review the baseline syntax line by line in harmony with semantic definition work taking place in the United Nations and to make whatever modifications are necessary in light of experience with other syntaxes:
X12 and EDIFACT
RosettaNet
OAG
cXML
Bolero
et cetera
The details of this process need to be defined in committee.
UN/CEFACT (UN/EDIFACT)
International participation and credibility
Links to major corporate users
Links to international organizations (WTO, WCO, IMF ...)
But subject to delays and political problems
OASIS
Nonprofit corporation founded in 1993
Represents XML vendors and users
Excellent, fast, fair committee process
Any interested person can join
[Status as of 25 July 2001]
Sun, Commerce One, SAP, and other software vendors have formed a UBL Organizing Committee
We are attempting to form a working group in UN/CEFACT to carry out the work of standardization
We are hoping to begin work after the UN/EDIFACT working group meeting in September
[Status as of 3 October 2001]
Pending a reorganization in progress in UN/CEFACT, the UBL work will take place in an OASIS technical committee
The first UBL TC meeting is scheduled for the week of 29 October through 1 November 2001
For more details: