What is XML? Extensible Markup Language A markup language: This is Italic becomes This is Italic Subset of ISO SGML More Features than HTML - Extensible A meta-language
XML - A Family of Recommendations XML 1.0 - Feb 1998 –Base language –Document Type Definition (DTD) XSL for Display - In Progress XSLT 1.0 - Transformations - 11/1999 XHTML 1.0 - HTML in XML - 1/2000 Schema for Data Definition - In Progress
The Vision of XML XML Document XSL Stylesheet XSL Stylesheet DTD Printed Document Web Page Business Application
XML for e-Transactions and impact on Architecture
Why XML for e-Transactions? Define tags for data Rawlins Many tools for software developers Can use mature file-oriented technologies XML Based Architecture
XML e-Transaction Conclusions In contrast to traditional EDI: –More pervasive –More modular –Simpler Transformations will still be required Subsystem integration via XML Business standards will be needed
X12 Analysis: Short Tag Names Target – application-to-application interactions Expectation – Short tag names require less communications bandwidth Issue – The shorter the tag, the more cryptic, and still need a source from which to derive semantic intent
X12 Analysis: Long Tag Names Target - application-to-human interactions Expectation - with long enough tag names, user will understand what is expected Issue - How long must the tag be to describe semantic intent?
X12 Tag Length Resolution Core Issue – Semantic Intent?
X12 Conclusions Mechanical generation of XML tag names not a desirable solution Semantic tags would require considerable effort Many issues remain the same Contribute to and support the ebXML efforts
ebXML Approach Adopt existing specifications where appropriate Utilize what worked in EDI Understand underlying reasons for failure in EDI Bridge needs for application-to-application versus application-to-human
What works in EDI? Modularity Grouping based on shared physical structure 50% Groupings based purpose in message 50% Business Document level Messaging 70%
What doesnt work in EDI Dependence on Implementation guides Implied Business Process Content and Presentation Intertwined One size fits all philosophy
Business Sector Specific Content for each BP function Subclassing a base business process, for a Sector application Sector Business Process Content, identified in terms of function or role in message ebXML Illustrated: Context impact on Core Components Core Data
Implementation Check List Trading Partner Readiness/Request Mature W3C Specifications Tools to support W3C specifications XML Business Standards Technical Infrastructure Standards Bodies Product Support
Near Term Strategies X12 still more appropriate for some things Applications with existing standards –Dont wait for XML, use current EDI –3 - 5 yr. ROI probably OK, depending on app. New applications may start with XML Preparation for XML –Target non-EDI partners –Start modeling
Web Exchanges - The Wild Card Mostly for procurement Few are integrated with business applications Could replace rip & read EDI What kinds of data will be exchanged? Will business models really change?
Summary You dont get something for nothing! –Shortcutting standards increases implementation costs –Win-win and not just cost transfer XML technology not ready for prime time, but rapidly evolving
Contact Information Lisa Shreve, Syscom Strategies, Inc. –firstname.lastname@example.org Mike Rawlins, Rawlins EDI Consulting –email@example.com –www.metronet.com/~rawlins