EDI Faults

1

• Premise
• EDIs Weak Points
• 14 XML/edi Advantages
• General Benefits of XML
• Summary

Eric Okin Assistant Deputy Director Policy and
Architecture DFAS - DTB
Mike Lubash XML Team Leader DoD Finance and
Accounting Namespace Manager
2
Premise
ANSI X12 Electronic Data Interchange (EDI)
restrictive scope, door-to-door, difficult to
repurpose, not real-time, not for web, long
turn-around to add to standard, messages are
extensible thus a decline in usage. Very few,
if any, new EDI Transaction Sets will be added to
X12/EDIFACT... Since introduction of the idea
of exchanging business documents by the XML/edi
Group in 1997, an unprecedented number of
exchange definitions and uses have been developed
and a plethora of tools have appeared on the
market.
3
EDI Weak Points - Specific to Where XML Plays

• Limited Internet support IETF
• Version control whole standard rolls
• Maintaining and updating the standards
  dictionary vs UIDs
• Cost of Implementation steep on-ramp
• Time to Implement discovery and exchange of
  metadata
• No provision for process and information
  exchange, data only
• Hurdles for near real-time
• Semantics
• Qualifiers overloaded
• ZZ never standardized
• Rationale not kept
• Structure not self-describing
• Grouping end tag processing
• Limited indentures HL loops

4
EDI Weak Points - Specific to Where XML Plays
1. Built with Internet in mind 2. At domain
component level 3. Dynamic registry with UIDs
(dfas.info) 4. Low cost tools on the market 5.
Collaborative Profiles / Agreements 6. Web
Services are verbs also there exists workflow
and BP schemas 7. Dynamic Registry
(dfas.info) 8. XML is self-keying into
Registry 9. with end tags 10. Unlimited
indentures...

• Limited Internet support IETF
• Version control whole standard rolls
• Maintaining and updating the standards
  dictionary vs UIDs
• Cost of Implementation steep on-ramp
• Time to Implement discovery and exchange of
  metadata
• No provision for process and information
  exchange, data only
• Hurdles for near real-time
• Semantics
• Qualifiers overloaded
• ZZ never standardized
• Rationale not kept
• Structure not self-describing
• Grouping end tag processing
• Limited indentures HL loops

5
10. Unlimited Indentures - Higher Granularity
EDI
XML
Very Rich Unlimited Number of Levels
Restricted 3-4 levels only
Can be extended via complex Hierarchical
Looping (HL) mechanism
6
11. Lowers the Bar Easy To Read
lt?xml version"1.0" encoding"UTF-8" ?gt -
ltSGLAccountsgt ltReportingDeptgtDept. of the
Armylt/ReportingDeptgt ltAccount Number"1010"
NBI"D"gt34507584472lt/Accountgt ltAccount
Number"1110" NBI"D"gt436988lt/Accountgt
ltAccount Number"1120" NBI"D"gt12112353lt/Accountgt
ltAccount Number"1130" NBI"D"gt0lt/Accountgt
ltAccount Number"1190" NBI"D"gt145618807lt/Accou
ntgt ltAccount Number"1195"
NBI"D"gt0lt/Accountgt ltAccount Number"1200"
NBI"D"gt57747632lt/Accountgt ltAccount
Number"1310" NBI"D"gt688542592lt/Accountgt
. .
Back ARMY 1010D34,507,584,4721110D436,988112
0D12,112,353 1130D01190D145,618,8071195D01200D57
,747,632 1310D688,542,5921310D357,282,8391319C0
1319C-41,291,4191320D01320D01330D0
If XML
If Flat File...
If EDI...
7
12. Semantic Reference Machine Interchange
Fixed Width - Version 1
1010D34,507,584,4721110D436,9881120D12,112,353
1130D01190D145,618,8071195D01200D57,747,632324545
4721110D436,9881 0D357,282,8391319C0319C-41,291,41
91320D01320D013300 4721110D436,9881120D12,112
Delimited - - Version 2
1010D34507,58 4,4721110D4369881120D12112,
3531130D01190D145,618,8071195D012003233
D57,747,632324545 4721110D436,9881
0D357,282,8391319C0319C-1,291,4191320D013
20D0
If the sender adds one other field the receiving
program needs to be modified or the interchange
breaks
8
12. Semantic References Version 3 Segmented
(Cont)
EDI X12
New data elements dont break the exchange
To interpret the fields the EDI parser needs to
reference the EDI standards to field the data
STD
9
12. Semantic References v.4 Smart Delimiters
(Cont)
XML
ltDategt02/05/02lt/Dategt ltFromgtJerry.Hinton_at_dfas.millt
/Fromgt ltTogtAttendeesListlt/Togt ltCCgtMike.Lubash_at_dfas
.millt/CCgt ltBCCgtJerry.Hinton_at_aol.comlt/BCCgt ltSubject
gtThank Yoult/Subjectgt
The XML parser will not break if a field is added
to the exchange and knows immediately context of
the information -more later
10
12. Semantic References EDI Comparison
(Cont)
Declarative Processing ltFromgt ltTogt
ltSubjectgt
XML
DoA
ltDategt02/05/02lt/Dategt ltFromgtJerry.Hinton_at_dfas.millt
/Fromgt ltTogtAttendeesListlt/Togt ltCCgtMike.Lubash_at_dfas
.millt/CCgt ltBCCgtJerry.Hinton_at_aol.comlt/BCCgt ltSubject
gtThank Yoult/Subjectgt
DoB
DoC
11
13. Processing Reducing the Mapping Alternative

• want a data mart of transaction data, but all
  current options require further map development

• desire a capability for XML-based communications
  and processing, without requiring more map
  development

Option A Translation
Fielded (RDBMS)
EDI or XML Feed
Reports
Mapping
Query
Option B XML Store/Query
Reports
Simple Conversion
XML Indexed
Query
12
14. Context Views Multiple levels of
Resolution
give multiple views of the information.
High Resolution
Low Resolution
e.g. External Organization Viewpoint
External Organization Viewpoint

XML provides multiple views (high and low
resolutions) which allow for additional
constraints for processing where as RDBMS
tables, no views are not included in the
structure. With RDBMS tables, the relationships
give a hint as to the possible views
Internal Organization Viewpoint
13
Benefits of XML
Lowers the Bar - easy to read for business users
and technologist alike providing a common
ground for communicating information. Available
labor pool is large due to the fact that XML
parallels HTML education and XML doesnt require
large amounts of specific training to leverage.
Machines can easily parse XML and align with data
in a robust manner. Independence - from
Operating Systems, Applications, Databases,
Software Language, Presentation, etc. XSL
stylesheets describe how to render data on
different devices (monitors, printers, palm
pilots, WebTV, voice and agent interactions).
Universal Clipboard - implemented as
hierarchical nodal trees XML can accommodate
entity-relationships, freeform, and network data
representations. Any application can validate
information prior to internal processing. With
XML, all nodes can use the same methods for
simplifying and automating processes.
End-tagging and consistent syntax enables enhance
detection of incomplete information packages.
Granularity - XML tagging provides
high-resolution access to data enabling
context-based searching and delta updates.
Contextual information improve the ability to
retrieve relevant information from total pool of
information. eXtensible - domain-specific
vocabularies, that enable tag names to be
specific business needs of a community (e.g.,
finance and accounting, human resources). Need
not be limited to standard transactions, and
many initiatives which to choose. Semantic
References - minimal prior knowledge of sender
application is necessary to process information.
Not positional or delimiter defined, thus
allowing flexible packaging based on business
needs. Context Views any application can
extract and separate information it needs to
satisfy business functions from other
facilitation types of information (e.g., routing,
security, archiving). Users (or applications)
can on-demand select data views (e.g., one record
or all, sort by different attributes, various
details) based on business needs/rules.
14
Summary
XML/edi