Reasoning About Inconsistencies in Natural Language Requirements

1
Reasoning About Inconsistencies in Natural
Language Requirements

• -VIVCENZO GERVASI
• -DIDAR ZOWGHI
• 
  BY
• 
  - ROHINI RAYI(2416209)

2
OUT LINE

• INTRODUCTION
• STRONG POINTS AND CONCENTRATION OF THE PAPER
• MAJOR CONTRIBUTIONS OF PAPER
• CARLTOOL
• BRINGING LOGIC TO USER
• DISCUSSION AND COMPARISON
• CONCLUSIONS AND FUTURE ADVANCEMENTS

3
SOFTWARE LIFE-CYCLE

• SYSTEM AND SOFTWARE REQUIREMENTS SPECIFICATIONS
• DESIGN
• IMPLEMENTATION
• INTEGRATION TESTING
• OPERATION MAINTENANCE

4
REQUIREMENT ENGINEERING???

• THE BRANCH CONCERNED WITH REAL WORLD GOALS
• RELATION BETWEEN SPECIFICATIONS AND SYSTEM
  BEHAVIOR

5
CAUSES OF SYSTEM FAILURE

• POOR REQUIREMENT MANAGEMENT
• INCORRECT SPECIFICATIONS
• INEFFECTIVE REQUIREMENTS MANAGEMENT

6
INCONSISTENCY

• MAIN CLASS OF DEFECT
• WHY IT OCCURS?
• CONFLICTING GOALS
• UNCOORDINATED CHANGES INTRODUCED

7
WAYS TO RESOLVE INCONSISTENCY

• TREAT IT AS AN ERROR
• USEFUL REASONING
• TYPES OF INCONSISTENCIES
• EXPLICIT
• IMPLICIT

8
STRONG POINTS AND CONCENTRATION OF THE PAPER
9
LOGICAL CONTRADICTION

• FACT AND ITS NEGATION FROM THE SAME SPECIFICATION
• CAUSES
• INFEASIBLE REQUIREMENTS
• DISAGREEMENT AMONG STAKEHOLDERS
• RECENT REQUIREMENTS CONFLICTING WITH PREVIOUS

10
PROPOSITION LOGIC

• SIMPLEST FORM
• EXPRESSIVE POWER LIMITED
• BEST FOR LARGE SET PROBLEMS
• SOUND AND COMPLETE
• GUARANTEED TO TERMINATE

11
NATURAL LANGUAGE???

• DISADVANTAGES
• LACKS FORMAL SEMANTICS
• MORE AMBIQUITY
• ADVANTAGES
• 95 OF REQ. DOCUMENTS
• ENCOURAGES EXPRESSION AND EXPERIMENTATION

12
MAJOR CONTRIBUTIONS
13
ARTICLE PRESENTS

• A formal framework for identifying, analyzing,
  and managing inconsistency in requirements
  specifications
• A parsing technique and a translation schema
• A prototype tool, called CARL

14
OVER VIEW
15
FORMAL MODEL DRAWS ITS RESULTS FROM

• DEFAULT REASONING
• FIRM
• DEFEASIBLE
• DISCARDED
• BELIEF REVISION
• COMPETENCE THEORY
• PERFORMANCE THEORY

16
THEORIST FRAMEWORKMODEL

• SPECIFICATION
• FACTS
• HYPOTHESES
• CONSTRAINTS

17
EVALUATION

• REVISION
• CONTRACTION

18
CARL TOOL

• MAINTAINS
• RELATIONAL DATABASE
• A SET OF BOUNDED-DEPTH, IN-MEMORY TREES

19
PARSING PROCESS

• PREPROCESSING STAGE
• TOKENIZATION
• MORPHOSYSNTACTIC ANALYSIS
• TYPOGRAPHICAL ADJUSTMENTS
• APPLYING FUZZY PARSING RULES
• BACKTRACKING
• HEURISTIC OPTIMIZATION STRATERIES
• TRANSLATION INTO LOGICAL FORMULAE

20
DISCUSSION AND COMPARISON
21
PROBLEMS SOLUTIONS

• PROPOSITION LOGIC NOT VERY EXPRESSIVE
• Sol RESTRICT THE LANGUAGE
• FINITE DOMAIN
• SCALABILITY
• Sol PARSING AND TRANSLATION, ONE AT A TIME.
• BOUND BY SIZE OF THE NL.

22
CONCLUSIONS AND FUTURE ADVANCEMENTS

• INTEGRATING CARL WITH REQUIREMENTS MANAGEMENT
• RECORDS ALL APPLICATIONS
• AUTOMATIC ERROR DETECTION
• COMBINES EXPRESSIVENESS OF NL WITH PRECISENESS,
  RIGOR, AND FORMALITY OF LOGIC.

23
QUESTIONS???