Forgot your password?
Abstract as submitted
Document Actions
This is the abstract we submitted to SPIP
Free and Open Source Software Project Success: Three new and appropriate measures
Kevin Crowston et al
School of Information Studies, Syracuse University
Email: {crowston, ...}@syr.edu
Information systems success is one of the most widely used dependent variables in information systems research. Research on Free (Libré) and Open Source software (FLOSS) suffers from a failure to appropriately conceptualize this important concept. In this paper, we reconsider what success means within a FLOSS context. We first review existing models of IS success and asses them for their usefulness, practicality and fit. Then, drawing on our model of group effectiveness in the FLOSS development process, we present three new measures that can be used, when appropriate and as part of a portfolio, to assess success in FLOSS and explain why they are particularly appropriate to the context of FLOSS.
We developed these new measures based on a review of the Information Systems literature, our theoretical model of group effectiveness in the FLOSS development process and an analysis of the opinions of FLOSS developers gathered through interviews and focus groups. Our three new measures are: the extent to which code is re-used, the project’s effectiveness in team building and the effectiveness of the project at coping with software faults.
The extent to which a project’s code is re-used is an important aspect of software use in FLOSS and reflects a systemic difference between the proprietary and the free and open software development communities. The ecosystem of FLOSS means that re-use can be measured by analyzing the network of project dependencies in software distributions.
Team building is important because FLOSS development relies on volunteer and collaborative labor, without which projects cannot continue or grow. This labor includes developers but also active users, neither of which are a typical responsibility of proprietary software development teams. To capture this we measured the rate at which projects attract and retain both developers and active users.
Our third measure is linked to our second; FLOSS advocates claim that FLOSS development is particularly effective because “many eyes make all bugs shallow” (ESR). We therefore measured, using event history analysis, the speed with which projects are able to identify and fix bugs.
Finally we report our efforts to validate these measures by assessing their construct validity, and their performance and practicality in pilot studies using data from three FLOSS repositories (Sourceforge, Apache and GNU Savannah) and two Linux distributions (Debian and Gentoo).
We conclude by considering the relevance of our measures to practitioners, both to FLOSS teams themselves and to companies considering “going open” or otherwise engaging with the FLOSS community.
Kevin Crowston et al
School of Information Studies, Syracuse University
Email: {crowston, ...}@syr.edu
Information systems success is one of the most widely used dependent variables in information systems research. Research on Free (Libré) and Open Source software (FLOSS) suffers from a failure to appropriately conceptualize this important concept. In this paper, we reconsider what success means within a FLOSS context. We first review existing models of IS success and asses them for their usefulness, practicality and fit. Then, drawing on our model of group effectiveness in the FLOSS development process, we present three new measures that can be used, when appropriate and as part of a portfolio, to assess success in FLOSS and explain why they are particularly appropriate to the context of FLOSS.
We developed these new measures based on a review of the Information Systems literature, our theoretical model of group effectiveness in the FLOSS development process and an analysis of the opinions of FLOSS developers gathered through interviews and focus groups. Our three new measures are: the extent to which code is re-used, the project’s effectiveness in team building and the effectiveness of the project at coping with software faults.
The extent to which a project’s code is re-used is an important aspect of software use in FLOSS and reflects a systemic difference between the proprietary and the free and open software development communities. The ecosystem of FLOSS means that re-use can be measured by analyzing the network of project dependencies in software distributions.
Team building is important because FLOSS development relies on volunteer and collaborative labor, without which projects cannot continue or grow. This labor includes developers but also active users, neither of which are a typical responsibility of proprietary software development teams. To capture this we measured the rate at which projects attract and retain both developers and active users.
Our third measure is linked to our second; FLOSS advocates claim that FLOSS development is particularly effective because “many eyes make all bugs shallow” (ESR). We therefore measured, using event history analysis, the speed with which projects are able to identify and fix bugs.
Finally we report our efforts to validate these measures by assessing their construct validity, and their performance and practicality in pilot studies using data from three FLOSS repositories (Sourceforge, Apache and GNU Savannah) and two Linux distributions (Debian and Gentoo).
We conclude by considering the relevance of our measures to practitioners, both to FLOSS teams themselves and to companies considering “going open” or otherwise engaging with the FLOSS community.
