Developing a knowledge-based perspective on coordination: The case of global software projects

Author(s):  
Julia Kotlarsky ◽  
Paul C. van Fenema ◽  
Leslie P. Willcocks
2015 ◽  
Vol 4 (2) ◽  
pp. 18 ◽  
Author(s):  
Georg Von Krogh ◽  
Nina Geilinger

Valve is a very interesting case study! The company shares many features with open source software projects. In Valve, as in the open source world, the focus is on creating advanced technologies and letting developers self-select projects and tasks. Self-selection seems to work particularly well in an environment where technology development itself has a coordinating function. In software development, people have a strong bond of common knowledge based on design philosophies and approaches, language and terminology, and engineering tasks. In most cases, they also share a similar educational background and/or development experience that helps them implement creative ideas in physical machines and graphic environments. Developers rarely rely on detailed instructions from higher-level managers because they already have the implicit knowledge of what needs to be done. When inconsistent views about development come to the forefront, they are best reconciled where the knowledge resides: with the experts. Moreover, it appears that Valve has uncovered how developers’ self-selection of projects and tasks can be a powerful motivator to unleash their creativity.


Author(s):  
Miguel Ángel Rodríguez-García ◽  
Rafael Valencia-García ◽  
Gema Alcaraz-Mármol ◽  
César Carralero

Author(s):  
Naeem Seliya ◽  
Taghi M. Khoshgoftaar

In machine learning the problem of limited data for supervised learning is a challenging problem with practical applications. We address a similar problem in the context of software quality modeling. Knowledge- based software engineering includes the use of quantitative software quality estimation models. Such models are trained using apriori software quality knowledge in the form of software metrics and defect data of previously developed software projects. However, various practical issues limit the availability of defect data for all modules in the training data. We present two solutions to the problem of software quality modeling when a limited number of training modules have known defect data. The proposed solutions are a semisupervised clustering with expert input scheme and a semisupervised classification approach with the expectation-maximization algorithm. Software measurement datasets obtained from multiple NASA software projects are used in our empirical investigation. The software quality knowledge learnt during the semisupervised learning processes provided good generalization performances for multiple test datasets. In addition, both solutions provided better predictions compared to a supervised learner trained on the initial labeled dataset.


2008 ◽  
Vol 45 (2) ◽  
pp. 96-108 ◽  
Author(s):  
Julia Kotlarsky ◽  
Paul C. van Fenema ◽  
Leslie P. Willcocks

10.5772/16377 ◽  
2011 ◽  
Author(s):  
Pasquale Ardimento ◽  
Nicola Boffoli ◽  
Danilo Caivano ◽  
Marta Cimitile

2009 ◽  
pp. 2714-2727
Author(s):  
Naeem Seliya ◽  
Taghi M. Khoshgoftaar

In machine learning the problem of limited data for supervised learning is a challenging problem with practical applications. We address a similar problem in the context of software quality modeling. Knowledge-based software engineering includes the use of quantitative software quality estimation models. Such models are trained using apriori software quality knowledge in the form of software metrics and defect data of previously developed software projects. However, various practical issues limit the availability of defect data for all modules in the training data. We present two solutions to the problem of software quality modeling when a limited number of training modules have known defect data. The proposed solutions are a semisupervised clustering with expert input scheme and a semisupervised classification approach with the expectation-maximization algorithm. Software measurement datasets obtained from multiple NASA software projects are used in our empirical investigation. The software quality knowledge learnt during the semisupervised learning processes provided good generalization performances for multiple test datasets. In addition, both solutions provided better predictions compared to a supervised learner trained on the initial labeled dataset.


Sign in / Sign up

Export Citation Format

Share Document