Business Maps Topic Maps Go B2B - PowerPoint PPT Presentation

1 / 38
About This Presentation
Title:

Business Maps Topic Maps Go B2B

Description:

marc_at_marcdegraauw.com. 1. B2B Interoperability. 2. Current Solutions to Interoperability ... marc_at_marcdegraauw.com. Type and token ... – PowerPoint PPT presentation

Number of Views:60
Avg rating:3.0/5.0
Slides: 39
Provided by: marcde8
Category:
Tags: b2b | business | maps | marc | topic

less

Transcript and Presenter's Notes

Title: Business Maps Topic Maps Go B2B


1
Business MapsTopic Maps Go B2B!
  • Marc de Graauw
  • XML Europe 2002, May 23

2
1. B2B Interoperability 2. Current Solutions to
Interoperability3. Problems with Current
Solutions 4. The Mapping Model 5. Business Maps
- a Proposed Solution6. Conclusions
3
B2B Interoperability - the problem
  • Abundance of vocabularies
  • xCBL, FinXML, FpML, ebXML, X.12, EDIFACT,
    legacy, in-house, proprietary
  • Business partners use different vocabularies
  • No quick solution
  • Interoperability is a HUGE problem

4
Type and token
  • token items
  • ltamount currencyUSDgt200.00lt/amountgt
  • USD20000
  • type items
  • amount is the total amount due ex. VAT
  • amount is in USD, EUR, JPY or IDR
  • amount has a maximum of two decimals

5
Type and token
  • this paper is about type-to-type mappings
  • linguistics type - token
  • IT class - instance
  • mathematics set - element

6
Type and token
  • IT / mathematics enumerable set (class) of
    elements (instances)
  • natural language type / token does not
    presuppose enumerability
  • compare 1, 2, 3 with love
  • clash of cultures
  • computers use class / instance
  • data definitions use natural language

7
What is money?
  • For me Euro
  • Was for me Dutch guilder
  • For you maybe dollar, yen, pound
  • Historian ducats, sestertii
  • Anthropologist shells, beads
  • Economist M0, M1, M2, M3, M4
  • Accountant VAT, taxes, EBITDA?

8
1. B2B Interoperability 2. Current Solutions to
Interoperability3. Problems with Current
Solutions 4. The Mapping Model 5. Business Maps
- a Proposed Solution6. Conclusions
9
1. The naive approach
  • Make a new vocabulary
  • Make it real real real good!
  • Make everybody use it!!!
  • ...multiple vocabularies are a fact of life

10
2. Adding meta-information
  • ebXML Core Components

11
2. Adding meta-information
  • Assemble Core Components in Business Document
  • Core Components dont have semantics
  • Use Context Rules to add information
  • Region, Industry, Business Process, Product,
    Official Constraints, Role, Temporal, Information
    Structural Context, Application Processing,
    Service Level, Business Purpose, Virtual
    Marketplace, Contractual

12
3. Public libraries of unique IDs
  • Use Published Subject Identifiers
  • Informal example
  • Topic 'United States of America' PSI US
  • Topic 'Verenigde Staten van America' PSI US

13
PSIs
mao
lenin
stalin
PSIs
topics
Vladimir Ilyich Ulyanov
Stalin
Joseph V. Stalin
Mao Ze Dong
Vladimir Ilich Lenin
Lenin
Mao Tse Tung
Iosif Vissarionovich Dzhugashvili
14
3. Public libraries of unique IDs
  • Good solution if
  • problem space with clearly delimited entities
  • widely accepted public library
  • ISO country codes, currency codes
  • Social Security Numbers
  • ZIP codes

15
4. Automatic Mapping
(design time)
biz document type 1
biz document type 2
mapping
generate
biz document token A (of type 1)
biz document token B (of type 2)
mapping
(run time)
16
4. Automatic Mapping
  • formal semantic notation for
  • type documents
  • type items
  • type-to-type mappings
  • complete (all relevant information from natural
    language is retained)
  • precise (only information which comes from
    natural language definitions)

17
1. B2B Interoperability 2. Current Solutions to
Interoperability3. Problems with Current
Solutions 4. The Mapping Model 5. Business Maps
- a Proposed Solution6. Conclusions
18
Trading Partner 1 CustomerAdress
Trading Partner 2 invoice_adress
Sales DB
Invoice DB
aaaa, bbbb, cccc pppp, qqqq, rrrr xxxx, yyyy, zzzz
19
True Equivalence?
  • When our partner does our billing
  • CustomerAdress ? invoice_adress
  • Not
  • CustomerAdress invoice_adress
  • That implies
  • CustomerAdress ? invoice_adress
  • We often dont establish equivalence but
    subclass/superclass relations

20
Problems
  • PSIs dont work without true equivalence
  • Automatic mapping is still in the lab
  • Context-based automatic mapping needs complete
    information
  • In reality there is often only partial
    information

21
What to do ?
  • Do we need new mappings between every two
    ontologies for every business process?
  • No knowledge in mappings is reusable - by human
    B2B analysts
  • So we need to capture the knowledge in mappings
    for reuse

22
1. B2B Interoperability 2. Current Solutions to
Interoperability3. Problems with Current
Solutions 4. The Mapping Model 5. Business Maps
- a Proposed Solution6. Conclusions
23
Trading Partner 1
Trading Partner 2
Sales DB
Invoice DB
billing
Invoice Mapping
CustomerName ? company_name CustomerAdress ?
invoice_adress TotalSales ? amount_due
Mailing DB
Sales DB
mailings
24
The Mapping Model
  • token-to-token mappings
  • XSLT
  • programming language
  • natural language
  • type-to-type mappings
  • no standard
  • (structured natural language)

25
Sample type-to-type mapping(xCBL 3.5 to X12)
26
(No Transcript)
27
1. B2B Interoperability 2. Current Solutions to
Interoperability3. Problems with Current
Solutions 4. The Mapping Model 5. Business Maps
- a Proposed Solution6. Conclusions
28
Business Maps
  • Keep data definitions external
  • no redundancy
  • external format is irrelevant (UML, TM, RDF,
    plain text, proprietary tools)
  • Topic Maps fit Mapping Model nicely
  • Topic Maps can provide portable, reusable
    mappings
  • Merging facilitates reusability

29
(No Transcript)
30
Position of mappings
Vocabularies
itm.xtm
context.xtm
Bizwords!
bizwords.xtm
GBL
gbl.xtm
businessmap1.xtm
businessmap2.xtm
inhouse
data dictionary
dd.xtm
31
Topic types
  • document
  • a business document, i.e. "Order"
  • item
  • an item in a document, i.e. "order date"
  • domain
  • a data type, i.e. "date"

32
Occurrence types
  • definition
  • a data definition as stored in a vocabulary
  • datatype
  • a definition of a data type as stored in a
    vocabulary

33
Association types
  • composition
  • a document is composed of items, i.e. "Order" is
    composed of "order date", "adress" et cetera
  • document_mapping
  • a document maps onto another, i.e. "Order" maps
    onto "Purchase Order"

34
Association types
  • item_mapping
  • an item maps onto another item, i.e. "order date"
    maps onto "purchase order date"
  • unidirectional
  • one-way item mapping, the reverse mapping is not
    allowed
  • to_yyyymmdd
  • date conversion to YYYYMMDD format

35
Role types
  • A composition has roles
  • containing-document, optional-item, required-item
  • An item_mapping has roles
  • source_item, destination_item
  • An document_mapping has roles
  • source_document, destination_document

36
1. B2B Interoperability 2. Current Solutions to
Interoperability3. Problems with Current
Solutions 4. The Mapping Model 5. Business Maps
- a Proposed Solution6. Conclusions
37
Conclusions
  • In B2B mapping, we often only have partial
    information
  • We need to capture the knowledge in mappings for
    reuse
  • Topic Maps provide an excellent vehicle to make
    such Business Maps

38
  • www.marcdegraauw.com/itm
Write a Comment
User Comments (0)
About PowerShow.com