Collaboration between Software Developers and the Impact of Proximity
1 like328 views
This research explores how paid software developers collaborate in fluid organizations using proximity theory across five dimensions: cognitive, organizational, social, institutional, and geographical. A case study of the Linux kernel community reveals that proximity significantly influences collaboration ties, with preliminary results supporting qualitative findings from developer interviews. Further research is suggested to validate these findings beyond the dataset period and context.
Collaboration between Software Developers and the Impact of Proximity
- 1. Research Overview The research uses five dimensions of proximity theory to explore this question: “How do participants, who are paid by firms, collaborate within a fluid organization?” Despite increased participation from paid software developers, little research has been conducted to investigate collaboration as it relates contributors who are employed by firms to work within a fluid organization. Research Setting Linux Kernel Community Case Study1: • Open source software • Over 85% of contributors paid • Neutral: competing companies • 19M lines of code • 11K developers • 1200 organisations References 1. Corbet, J., Kroah-Hartman, G. & McPherson, A., 2015. Linux Kernel Development: How Fast is it Going, Who is Doing It, What Are They Doing and Who is Sponsoring the Work, Available at: http://www.linuxfoundation.org/publications/linux-foundation/ who-writes-linux-2015. 2. March, J.G. & Simon, H.A., 1993. Organizations Second Ed., Malden, MA: Blackwell. 3. Dobusch, L. & Schoeneborn, D., 2015. Fluidity, Identity, and Organizationality: The Communicative Constitution of Anonymous. Journal of Management Studies, 52(8), pp.1005–1035. 4. Glance, N.S. & Huberman, B.A., 1994. Social dilemmas and fluid organizations, Hillsdale, NJ: Lawrence Erlbaum. 5. Balland, P.A., 2012. Proximity and the Evolution of Collaboration Networks: Evidence from Research and Development Projects within the Global Navigation Satellite System (GNSS) Industry. Regional Studies, 46(6), pp.741–756. 6. Crescenzi, R., Nathan, M. & Rodríguez-Pose, A., 2016. Do inventors talk to strangers? On proximity and collaborative knowledge creation. Research Policy, 45(1), pp.177– 194. 7. Knoben, J. & Oerlemans, L. a G., 2006. Proximity and inter-organizational collaboration: A literature review. International Journal of Management Reviews, 8(2), pp.71–89. 8. Cantner, U. & Graf, H., 2006. The network of innovators in Jena: An application of social network analysis. Research Policy, 35(4), pp.463–480. 9. Boschma, R., 2005. Proximity and Innovation: A Critical Assessment. Regional Studies, 39(1), pp. 61–74. 10. Butts, C.T., 2008. A relational event framework for social action. Sociological Methodology, 38(1), pp.155-200. 11. Quintane, E., Pattison, P.E., Robins, G.L. and Mol, J.M., 2013. Short-and long-term stability in organizational networks: Temporal structures of project teams. Social Networks, 35(4), pp.528-540. 12. Opsahl, T. and Hogan, B., 2011. Modeling the evolution of continuously-observed networks: Communication in a Facebook-like community. arXiv preprint arXiv: 1010.2141. Method Relational Event Framework • Predicting events in an ordinal sequence is product of multinomial likelihoods.10 • Ordinal model estimated using Multinomial Conditional Logistic Regression, specifically Cox regression estimated using MLE.11 • Using clogit in R, which is based on coxph. • Realized event compared to 3 randomly sampled possible events.12 • 10 day moving window. Background March and Simon2 define organizations as systems for coordinating activities between individuals to facilitate cooperation with a focus on supporting decision-making processes. The notion of organization can be expanded to include fluid organizations that emerge when people collaborate and make decisions within a community that is recognized by its collective identity.3 Collaboration between individuals occurs within these fluid organizations; however, collaboration within fluid organizations has been shown to reveal complex behavior with many dimensions.4 Proximity theory can been used to investigate various dimensions of collaboration5,6,7 and other complex topics related to collaboration, such as knowledge transfer and innovation.8,9 There are several approaches to proximity theory7, and this research uses five dimensions: cognitive, organizational, social, institutional and geographical.9 Collaboration between Software Developers and the Impact of Proximity Dawn M. Foster, Guido Conaldi, Riccardo De Vita Business School, Centre for Business Network Analysis Data Descriptive Statistics • Dataset: USB Mailing List (linux-usb) 2013-11-01 - 2015-11-01 • Messages (Events): 7799 in 3264 threads • Ties: based on Ego replying to a message from Alter • Actors: 882 (Egos: 691, Alters: 717) Variable Operationalization Proximity: • Geographic: time zone similarity (temporal geo prox) • Organizational: both work for same firm • Social prox: # of times dyad participated in same thread • Cognitive prox: contribute to same source code subsystems • Institutional prox: both employed by firms Dyadic-Level Covariates: • Is Maintainer: one or both are in leadership (maintainer) position • Is Committer: one or both have made code contributions • Alter Maintainer: Alter is in a leadership (maintainer) position Network-Level Covariates: • Transitive closure: num of x’s ego replied to where x has replied to alter • Cyclic closure: num of x’s alter replied to where x has replied to ego • Shared partnership in: same x replies to both ego and alter • Shared partnership out: ego and alter reply to messages by same x • Repeated events: number of times ego replied to messages by alter • Recency effect: 1/n with n as number of people alter emailed before ego • Participation shift: 1 if last person alter replied to on mailing list was ego xe a xe a e a e a a 1/3 1/2 1 xa e xe a XXXVII Sunbelt Conference 30 May 2017 – 4 June 2017 Beijing, China Preliminary Results • Proximity is relevant in explaining collaboration ties within a fluid organization. • Preliminary results are aligned with qualitative analysis from interviews with software developers in this setting. • Further Research: Expand beyond 2 years of data from one mailing list to see if the same results hold for other mailing lists. coef exp(coef) se(coef) org proximity 5.763e-01 1.779e+00 6.280e-02 *** social prox 3.369e+01 4.290e+14 1.047e+00 *** cognitive prox -4.620e-01 6.301e-01 1.237e-01 *** geo proximity 1.756e-01 1.192e+00 9.354e-02 . inst prox (corp)2.597e-01 1.297e+00 4.535e-02 *** is maintainer 5.128e-01 1.670e+00 1.167e-01 *** is committer 3.335e-01 1.396e+00 5.548e-02 *** alter maint -6.667e-01 5.134e-01 3.894e-01 . cyclic closure 1.685e+01 2.080e+07 7.209e-01 *** shared part in -3.263e+01 6.721e-15 1.020e+00 *** shared part out-2.713e+01 1.653e-12 1.095e+00 *** transitive clsr 1.060e+00 2.885e+00 5.555e-01 . repeated events 1.684e+01 2.051e+07 5.773e-01 *** recency effect 6.070e+00 4.326e+02 2.362e-01 *** particip shift -3.090e+00 4.550e-02 2.386e-01 *** --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1