Evaluating the Impact of UML Introduction Modeling on Software - - PowerPoint PPT Presentation

▶
evaluating the impact of uml
SMART_READER_LITE
LIVE PREVIEW

Evaluating the Impact of UML Introduction Modeling on Software - - PowerPoint PPT Presentation

Aug 11, 2023 204 likes •416 views

Evaluating the Impact of UML Introduction Modeling on Software Quality Research Design Case Study An Industrial Case Study Results Conclusions Ariadi Nugroho Michel R.V. Chaudron Modeling, is it worth it? Introduction Research Design

slide-1
SLIDE 1

Evaluating the Impact of UML Modeling on Software Quality

An Industrial Case Study

Introduction Research Design Case Study Results Conclusions

Ariadi Nugroho Michel R.V. Chaudron

slide-2
SLIDE 2

Introduction Research Design Case Study Results Conclusions

Modeling, is it worth it?

Well-thought solutions Better comprehension Improved Communication Reduced maint. effort Reduced reworks Reduced defects Improved productivity Cost-saving

slide-3
SLIDE 3

Introduction Research Design Case Study Results Conclusions

UML is used diversely: style, completeness, level of detail, formality...

But, this is often how people model...

slide-4
SLIDE 4

Introduction Research Design Case Study Results Conclusions

Exploring the effects of quality in modeling

  • Defects in UML models often remain undetected and

cause misinterpretation (Lange & Chaudron, ICSE 2006)

  • Higher level of detail in UML models increases

comprehension performance (IST Journal, 2009)

  • Higher level of detail in sequence diagrams corresponds

to lower defect density in software (MODELS 2008)

This study focuses on the use of UML and its effect on the quality of software

Does model quality matter?

slide-5
SLIDE 5

In modeling a system, some parts might have been left unmodeled for various reasons...

Introduction Research Design Case Study Results Conclusions

Implementation::Package A UML Model::Package A

do_A() aClass do_B() bClass do_C() cClass do_something() letterClass do_A() aClass do_B() bClass do_Z() zClass do_something() letterClass do_BB1() bb1Class do_BB2() bb2Class

  • Implementation may not be well-thought of or well-designed
  • Implementation may deviate from requirements

What may be the results of not modeling?

slide-6
SLIDE 6

“Do the modeled and not modeled classes have significantly different quality in the implementation?”

25 50 75 100 Modeled Not Modeled

Defect density

Introduction Research Design Case Study Results Conclusions

Research Question

slide-7
SLIDE 7

Introduction Research Design Case Study Results Conclusions

Hypothesis: Modeled classes have significantly lower defect density than those that are not modeled

Design of Study

No Modeling UML Modeling Main factor Confounding factors

Code Complexity Code Coupling

THE USE OF UML RESULTING IMPLEMENTATION

Defect Density Affects?

slide-8
SLIDE 8

Project characteristics

  • A web-based healthcare system
  • System modeling was done in Holland
  • Sixty percent of coding and testing were done in India
  • UML model was created using Rational XDE
  • Textual specifications exist

Project Size # of staffs Off-shoring Model Size Code size 20 man-years 25 people India 104 UCs 266 classes 341 SDs 152, 017

Introduction Research Design Case Study Results Conclusions

System under Study

slide-9
SLIDE 9

Introduction Research Design Case Study Results Conclusions

Approach

Payment
  • amount
  • cashTendered
Cash +authorized()
  • name
  • bankID
Check +authorized()
  • number
  • type
  • expDate
Credit +calcTax() +calcTotal() +calcTotalWeight()
  • date
  • status
Order +calcSubTotal() +calcWeight()
  • quantity
  • taxStatus
OrderDetail 1 1 1 * Preprocessor Parser Analyser Inheritance Relator Statistic Filter Statistic Calculator DB Creator DB Filler DB Checker SAAT.BAT

component C2 requires I2 requires I3 uses I2.g uses I3.h component C2 requires I2 requires I3 uses I2.g uses I3.h

defect density Modeled vs Unmodeled

13.75 27.50 41.25 55.00 Not Modeled Modeled

Collect implementation classes, UML model, and defect data Determine defect density of every implementation class Check whether an implementation class has been modeled in UML

Project Repository

Statistical tests

slide-10
SLIDE 10
  • The core part of the system contains 812 Java classes

Classes that were not modeled tend to have higher defect density than the modeled ones

Introduction Research Design Case Study Results Conclusions

21% 79%

Modeled Not Modeled

0.005 0.010 0.015 0.020

Not Modeled Modeled

Mean

Defect Density

Descriptive Statistics

slide-11
SLIDE 11

Class coupling and complexity might confound the effect of UML modeling on defect density

Introduction Research Design Case Study Results Conclusions

7.5 15.0 22.5 30.0 Not Modeled Modeled

Coupling Complexity

Mean

Descriptive Statistics

Modeled classes have significantly higher coupling and complexity

slide-12
SLIDE 12

Eliminating the effects of class complexity and coupling

Introduction Research Design Case Study Results Conclusions

CBO Total Variance in Defect Density Variance explained by the use of UML Unexplained Variance MCC

Controlling for Confounding Factors

slide-13
SLIDE 13

Introduction Research Design Case Study Results Conclusions Mean

11.25 22.50 33.75 45.00 Not Modeled (59) Modeled (37)

Coupling Complexity

Accounting for CBO Accounting for CBO & MCC

2 1

Implementation Classes (812 classes)

CBO 8 - 10 (113 classes) Random sampling (80% = 96 classes)

Mann-Whitney Test ANCOVA Test

Pairwise sampling

Controlling for Confounding Factors

slide-14
SLIDE 14

Result of Mann-Whitney Test

  • After controlling for the effect of class coupling

Not modeled classes have significantly higher defect density (p=0.003; 1-tailed)

Introduction Research Design Case Study Results Conclusions

13.75 27.50 41.25 55.00 Not Modeled Modeled

Defect Density (Mean Rank)

Testing the Hypothesis

slide-15
SLIDE 15

Result of ANCOVA Test

  • The use of UML remains a significant factor

that explains the variability of defect density in the implementation

Variables Sum of Squares df F Sig. The use of UML 0.002 1 6.825 0.010 Coupling 7.562E-8 1 0.000 0.987 Complexity 0.001 1 5.224 0.025

Introduction Research Design Case Study Results Conclusions

Testing the Hypothesis

slide-16
SLIDE 16

So,...

Introduction Research Design Case Study Results Conclusions

The hypothesis is confirmed: Modeled classes have significantly lower defect density than those that are not modeled.. after controlling for the effects of class complexity and coupling

Defect Density (Mean Rank)

0.005 0.010 0.015 0.020

Not Modeled Modeled

slide-17
SLIDE 17
  • The use of UML has a positive effect on the quality
  • f the system (i.e., reduced defect density)
  • Increasing model coverage might help reduce

incorrect implementation of systems

  • Deciding which classes to model is an open issue

Introduction Research Design Case Study Results Conclusions

Conclusions

slide-18
SLIDE 18

Introduction Research Design Case Study Results Conclusions

  • Replication using more industrial cases
  • Accounting for the effects of other confounding

factors

  • Performing similar studies in experimental

settings

What’s Next?

slide-19
SLIDE 19

Introduction Research Design Case Study Results Conclusions

Questions?