Upload
marco-wobben
View
167
Download
1
Embed Size (px)
DESCRIPTION
At Software Development Automation 2014, BCP Software presented CaseTalk to show how to build software using natural language.
Citation preview
From natural language to working software
SDA Conference -Amsterdam 2014 Speaker: Marco Wobben
Brief history• NIAM / Nijssen (‘70s)
• FCO-IM / G. Bakema, JP. Zwart, H vd. Lek (‘92)
• ORM / T. Halpin (’96)
• CaseTalk (2001)
FCO-IMFully Communication Oriented Information Modeling
• 100% conceptualization
• Redundancy free
• Unification: facts and objects
• Substitution: expressions and tuples
• Generic repository
Fact based modeling• Domain expert participation
• Improved model quality; Consolidate business knowledge
• Complex business domains • Enterprise data integration
Technology independence
• More complete requirements, faster • Cost reduction
Example
34-KK-NN
Communication
Gasoline Consumption "The car with license number 34-KK-NN drives 12.0km per litre."
34-KK-NN
ModelingGasoline Consumption "The car with license number 34-KK-NN drives 12.0km per litre."
FCO-IM Diagram
UML Diagram
ER Diagram
Database Script
Verbalizations------------------------------------------------------------------ Facts ------------------------------------------------------------------
Gasoline Consumption: "The car with license number 34-KK-NN drives 12.0km per litre."
------------------------------------------------------------------ Business Rules ------------------------------------------------------------------
Car: "Car is uniquely identified by License number." "Car must have a Car Fuel."
Distance: "Distance is uniquely identified by #KM." "Distance must have a Gasoline Consumption."
Gasoline Consumption: "Gasoline Consumption is uniquely identified by Car."
CaseTalk Demo
CaseTalk Diagram
CaseTalk Script
CaseTalk ERD
CaseTalk APP
CaseTalk
CaseTalk
DDL Scripts, ERD, XMI, UML Classes Data Warehouse, XML Schema, OWL/RDF, Prototyping, etc..
Verification Evaluation
Negotiation
Verbalization, Concepts, Vision
Educational
Customers
Project metrics• ProRail / Master Data Management
> 1000 fact types
• KLM / Passenger & Cargo> 1500 facttypes
• HAN University / Operational> 2500 fact types
• Erasmus MC / Data Warehouse> 3500 fact types