prezentation

22
“The impact of global dispersion on coordination, team performance and software quality – A systematic literature review” Anh Nguyen-Duc, Daniela S. Cruzes, Reidar Conradi

Upload: khush-bakhat

Post on 07-Aug-2015

57 views

Category:

Software


0 download

TRANSCRIPT

Page 1: Prezentation

“The impact of global dispersion on coordination, team performance and software

quality – A systematic literature review”

Anh Nguyen-Duc, Daniela S. Cruzes, Reidar Conradi

Page 2: Prezentation

Abstract: Context:

Global software development (GSD) contains different context setting dimensions, which are essential for effective teamwork and success of projects.

Objective:This paper summarizes empirical evidence on the impact of global dispersion dimensions on coordination, team performance and project outcomes.

Method:We performed a systematic literature review of 46 publications from 25 journals and 19 conference and workshop proceedings, which were published between 2001 and 2013.

Page 3: Prezentation

INTRODUCTION:• Software development and maintenance tasks are continually dispersing

globally for cost saving, time-to-market shortening, technology innovation and operational efficiency

• A large amount of empirical research has been devoted to understand general challenges.

managing collaboration getting requirements establishing trust communication and coordination engineering process in GSD.

Page 4: Prezentation

RESEARCH QUESTIONS:

• RQ1: How are dispersion dimensions defined and measured in GSD studies?

• RQ2: What are different coordination challenges that dispersion dimensions present to GSD project outcomes?

• RQ3: How does dispersion dimensions affect team performance in GSD project?

• RQ4: How does software quality affect software quality in GSD project?

Page 5: Prezentation

RESEARCH METHODOLOGY:

The complete review lasted for 23 months, including six main steps:1. Ad-hoc review2. Pilot search3. Systematic search & Extraction4. Additional Manual Search5. Quality Assessment6. Data Analysis

Page 6: Prezentation

REVIEW PROTOCOL:1. Search terms

Search string consisted of 3 parts: (Coordination Or synonyms) AND (Dispersion Or synonyms)

AND context.

• The ad-hoc review revealed that there was no distinguished use of terms ‘‘coordination,’’ ‘‘collaboration’’ and ‘‘cooperation’’

• Consequently, a list of related terms was identified: distributed, distribution, dispersed, dispersion, remote, offshoring, offshore,outsourcing, outsource, nearshore, and global.

Page 7: Prezentation

2. Search Database: Scopus, ISI Web of Science, IEEE Explore, Current Contents,

Kluwer Online, Computer Data-base, Science Direct, Springer Link, Inspec and ACM Digital Library

Page 8: Prezentation

3. Inclusion & Exclusion Criteria: We included papers that:

• Report studies in GSD context• Explore one of the global dispersion dimensions and relate them to either

team coordination or a specific type of project outcomes.

We excluded papers that:

• Investigate team performance or quality in GSD context, but not explicitly address the relationship between coordination or global dispersions and project outcomes.

• Investigate team coordination in other domains, such as business, education science and system engineering.

• Provide a conceptual framework or model about coordination that results from literature reviews or author’s thoughts without empirical validation.

Page 9: Prezentation

Measurement Scale:

• an issue is not mentioned at all (score 0),• little mentioned (score 1),• adequately addressed (score 2) • completely addressed (score 3).• average quality score equal or greater than 1

Page 10: Prezentation
Page 11: Prezentation

5. Data Analysis:

• A tailor thematic synthesis was conducted.• Identified the relevant codes from the data extraction form. • Selected the relevant piece of text and labeled them.• We applied C4.5 algorithm to build a classification tree

of impact direction based on these context factors

Page 12: Prezentation

RESULTS? ??

?

? ??

Page 13: Prezentation

RQ1: How are dispersion dimensions defined and measured in GSD studies?

Page 14: Prezentation

RQ2: What are different coordination challenges that dispersiondimensions present to GSD project outcomes?

DIMENSIONS CHALLANGES

Geographical dispersion

• decrease of communication frequency,• difficulty in finding relevant stakeholders

Temporal dispersion

• leads to time lag in communicating tasks• difficulty in scheduling meetings and• limited ability to communicate

Cultural dispersion

• largest issue is misinterpretation during all types of development activity: communication, requirement, elicitation, development, maintenance, and testing. • understanding of requirements• knowledge sharing between the client and vendor teams.• Organizations are not ready to change their internal structure and processes along with new partners.

Coordination dispersion

• harder to divide tasks and to assign responsibility across sites than in ahomogeneous environment.• difficult for team members in identifying roles and responsibilities.• create conflicts.

Page 15: Prezentation

Reference Geographical, cultural, temporal & coordination Dispersion

Cramton et al. (2005)

• surveyed 218 team members from 39 dispersed teams and showed that teams with geographically dispersed members report significantly less effective performance than collocated teams

Espinosa et al. (2007)

• geographic dispersion and team size have a negative impact on team performance. this impact is mediated by team familiarity.• Without team familiarity, geographically dispersed team members need to bridge the geographical distance to coordinate their work• making it more difficult to work together because of lacking presence awareness, frequent communication, and contextual reference.

RQ3: How does dispersion dimensions affect team performance inGSD project?

Page 16: Prezentation

References Cont….Gopal et al. (2011)

• surveyed 83 projects in 9 firms and found that temporal dispersion is positively associated with development speed.• The study stated that ‘‘follow-the-sun’’ practices allow managers to leverage 24 h work-day.• As time separation is larger, i.e., India and the US vs. India and Singapore, the team has more advantages in productivity.

Ramasubbu et al. (2008, 2011)

According to the authors,a ten-person project team configured as 8–2 between New YorkCity and Bangalore is more productive than 6–4 and 5–5 configurations

Ramasubbu et al. (2011)

• investigated 362 GSD projects from four organizations and showed thatgeographical distance leads to higher productivity.• collaboration between a pair of developers is delayed, in project level, firms gain benefits from being distributed

Page 17: Prezentation

RQ4: How does software quality affect software quality in GSDproject?

References Geographical dispersionBird et al. (2009) • Microsoft Windows Vista project and found that there is no

significantly correlation between different degree of geographical distribution and software quality

Bird et al. (2012) • Firefox and Eclipse and found that all geographical distributionmeasure increase number of failure at file level.• In Firefox case, distribution measures are more significant during prior to release than after. • Higher values of spatial and worldwide distribution were associated with more defects in both prior to both releases .• In the Eclipse case, there are also a mixed finding when all measures of geographic and organizational distribution increased failures, but the effects are not consistent across releases

Cataldo et al.(2009)

• investigated 562 components from GSD projects and foundthat an increase in distances among locations leads to an increaseof number of defects

Page 18: Prezentation

References Temporal dispersion

Cataldo et al. (2010)

• components that are developed by a team with a higher level of temporal dispersion exhibited lower levels of quality. • However, the authors did not find a significant relationship between temporal dispersion and software quality at the project level

Colazo et al. (2011)

• investigated 100 projects and found that a higher temporal dispersion is associated with fewer defects and subsequently with a higher quality of coding.• the positive effect of temporal boundary on quality is greater in more complex tasks.

Page 19: Prezentation

Work dispersion:

• a ten-person project team configured as 8–2 between New York City and Bangalore is more defect prone than 6–4 and 5–5 teams do

• Uneven distribution of work items, i.e. modification request, across development locations negatively impacted software quality.

Page 20: Prezentation

Summary:

Page 21: Prezentation

Future Work:(1) coordination challenges in inter-organizational projects,(2) Impact of processes and practices mismatches on project

outcomes, (3) evolution of coordination needs and mechanism over time and(4) impact of dispersion dimensions on open source project out-

comes. For GSD managers(4) the tradeoff between cost and benefits while dispersing tasks,(5) alignment impact of dispersion dimensions with individual and

organizational objectives,(6) coordination mechanisms as situational approaches (7) collocation of development activities of high quality demand

components

Page 22: Prezentation

THANK YOU FOR YOUR ATTENTION…