william krebs – ibm laurie williams, lucas layman, nc state

8
Quality through Software Engineering April, 2007 © 2007 IBM Corporation BH http://w3.ibm.com/ibm/presentations Extreme Programming Evaluation Framework XP:EF William Krebs – IBM Laurie Williams, Lucas Layman, NC State Bill Krebs Software Transformation Consultant IBM Corporate: QSE Durham, NC USA Phone: 919 486-0238 | T/L: 526-0238 [email protected]

Upload: zasha

Post on 06-Jan-2016

24 views

Category:

Documents


0 download

DESCRIPTION

Extreme Programming Evaluation Framework XP:EF. William Krebs – IBM Laurie Williams, Lucas Layman, NC State. Bill Krebs Software Transformation Consultant IBM Corporate: QSE Durham, NC USA Phone: 919 486-0238 | T/L: 526-0238 [email protected]. Evaluation Framework. Approach: - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: William Krebs – IBM Laurie Williams, Lucas Layman, NC State

Quality through Software Engineering

April, 2007 © 2007 IBM CorporationBH

Extreme Programming Evaluation FrameworkXP:EF

William Krebs – IBMLaurie Williams, Lucas Layman, NC State

Bill KrebsSoftware Transformation ConsultantIBM Corporate: QSEDurham, NC USAPhone: 919 486-0238 | T/L: [email protected]

Page 2: William Krebs – IBM Laurie Williams, Lucas Layman, NC State

2

Evaluation FrameworkEvaluation Framework

X P -c fC on text F ac to rs

X P -amA d h eren ce M etrics

X P -r mR esu lts M etrics

X P :E FX P E va lu a tion F ram ew ork

Goals:• Concrete• Easy to measure by small team• Unbiased

Approach:• Include objective• Include subjective

• Leading indicator• Teaches Team

• They cross check each other

Page 3: William Krebs – IBM Laurie Williams, Lucas Layman, NC State

3

XP:EF

| April 2007 © 2007 IBM CorporationBH

Context Factors (IBM Study)

Small team (7-10) Co-located Web development (toolkit) Supplier and customer

distributed (US and overseas)

Examined one release “old” (low XP) to the next “new” (more XP)

Page 4: William Krebs – IBM Laurie Williams, Lucas Layman, NC State

4

Planning PracticesPlanning Practices

1313 Stories AddedStories Added 1 1 Story RemovedStory Removed 60 60 Total StoriesTotal Stories 5 5 Month releaseMonth release 1 1 Week iterationsWeek iterations

+ + Shodan chartShodan chart

Shodan Survey

Planning

Cust Access

Short Rel

Standups

70 80 90 100

Page 5: William Krebs – IBM Laurie Williams, Lucas Layman, NC State

5

Testing PracticesTesting Practices

Shodan Survey

Cust Acceptnce

Auto Tests

Test First Design

40 60 80 100

46% Test Coverage46% Test Coverage (Alpha Set)(Alpha Set)

11% Test Frequency11% Test Frequency 0.45 Test to Story Ratio0.45 Test to Story Ratio 0.42 Test LOC / Src 0.42 Test LOC / Src

LOCLOC

Manual Customer Manual Customer Acceptance TestsAcceptance Tests

No Supplier No Supplier Acceptance TestsAcceptance Tests

+ Shodan at left+ Shodan at left

Page 6: William Krebs – IBM Laurie Williams, Lucas Layman, NC State

6

Coding PracticesCoding Practices

Pair Prog

Cont Integration

Collective Ownership

Coding Standards

Sustain Pace

Simple Design

Refactorng

Metaphor

40 60 80 100

48% Pairing Frequency48% Pairing Frequency

+ + Survey at leftSurvey at left

Be aware of pairing patterns:

Expert / ExpertExpert / NoviceNovice / Novice

Page 7: William Krebs – IBM Laurie Williams, Lucas Layman, NC State

7

IBM: XP-Outcome IBM: XP-Outcome MeasuresMeasures

XP Result Metric XP Result Metric Old Old New New CommentsComments

Response to Customer Response to Customer ChangeChange(Ratio (user stories in (Ratio (user stories in + out) /total)+ out) /total)

NANA 0.230.23 No basis for No basis for comparison, but comparison, but shows flexabilityshows flexability

Pre-release QualityPre-release Quality(test defects/KLOEC of (test defects/KLOEC of

code)code)

1.01.0 0.500.50 Improved!Improved!

Post-release QualityPost-release Quality(released defects/KLOEC (released defects/KLOEC

of code)of code)

1.01.0 0.610.61 Improved!Improved!

Productivity (stories / PM)Productivity (stories / PM)Relative KLOEC / PM Relative KLOEC / PM

1.01.01.01.0

1.341.341.71.7

Improved ?Improved ?Hard to measureHard to measure

Customer SatisfactionCustomer Satisfaction NANA High High GoodGood

Morale (via survey)Morale (via survey) 1.0 1.0 1.111.11 Significant? Other Significant? Other variables?variables?Moved from 57% to 72% of full use of all XP practices

Page 8: William Krebs – IBM Laurie Williams, Lucas Layman, NC State

8

XP:EF

| April 2007 © 2007 IBM CorporationBH

References

XP:EF spreadsheet– xpef_14.xls by William Krebs.

Concept paper– Laurie Williams, William Krebs, Lucas Layman, and Annie I. Antón,

“Toward a Framework for Evaluating Extreme Programming,” Proceedings of the 8th International Conference on Empirical Assessment in Software Engineering (EASE ’04), Edinburgh, Scotland, pp. 11-20

– http://www4.ncsu.edu/~lmlayma2/papers/WKL04.pdf