Revisiting my refocusing
November 4, 2008
I’m asking myself why I need to understand OSS to do my research, even though my data will come from open source software communities and projects.
Revisiting my refocusing
I had said recently that I was moving pretty slowly on my research and I wanted to revisit what it was that I was trying to accomplish in investigating open source software development and communities. But now I’m asking myself why I need to understand OSS to do my research, even though my data will come from open source software communities and projects. Other researchers don’t seem to worry too much about the special characteristics of open source while using OSS data in their research.
I’m afraid of biting off too much, and wasting precious time doing so. My advisor and other faculty in IS have said: “do the minimum amount of work to satisfy the requirements for your dissertation. It is not your magnum opus.” Why do I need to understand OSS to do my research? I qualify this by saying that I wasn’t going to do a bunch of research in organizational science and sociology about how traditional organizations work. To ground my original research question, I was, in fact, going to rely on my and my audience’s shared experiences of working in traditional organizations, plus a minimal amount of theoretical framework (Crozier and Friedburg, plus some reviews of power related literature from IS), but I wasn’t going to delve this deeply into it.
This is really a question about my research question. My original research question was something like:
I am interested in doing design research in the area of power networks in traditional organizations because I want to find out whether I can make power relationships visible to all members of the organization in order to show people that workers and organizations can be more effective when everyone understands the power network.
But then I changed my subject of inquiry from traditional firms to open source software development and communities. I think that the crux of why I ask myself “Should I be spending all this time learning about OSS?” is in how I phrase my revised research question.
Is it:
I am interested in doing design research in the area of power networks in open source software communities because I want to find out whether I can make them visible to the world in order to show people that OSS projects and communities can be more effective when everyone understands the power network.
or is it:
I am interested in doing design research in the area of power networks in open source software communities because I want to find out whether I can make them visible to the world in order to show people that workers and organizations can be more effective when everyone understands the power network.
The difference between both of these revised statements and the original one concerns a change of the subject of inquiry. I suppose that at any rate, no matter which revised question I go with, I need to understand the intrinsic and accidental characteristics of OSS and whether I need to do special things to accommodate them in my work. So I do need to do what I’m doing. I am assuming that OSS communities and development differ from traditional organizations and software development practices. From the research literature, I can see that there must be a difference, just from looking at research topics through the last ten years: first “will OSS be able to compete with firm-produced software?”; then “how does OSS work?” (still working on this one) and now “what does the success of OSS mean for the world?”
The difference in the two revised statements concerns generalizability vs. specificity. Do I want to focus my research and its relevance on OSS, or aim to make it more generally applicable? And if the latter, is this something that is reasonable as a Ph.D. research project? (Is any of it reasonable?)
I’m thinking that the additional question of whether those accommodations I may make compromise the generalizability of my work must come later, if at all. I think that eventually I want my work to be not limited to OSS contexts — in fact, I still want to do my original project and be able to eventually apply what I find to traditional organizations. But I’ll have to wait to see when and whether I can make that happen.
Addendum
I suppose that the other researchers who use OSS data in their research but ignore the OSS-ness of it are typically people looking to write software development support tools, or those looking at making better software reliability or estimation metrics. Community dynamics don’t play much or any part in such work, so they’re justified in using OSS data to produce results generalizable to all software development. The work I plan to do is intrinsically about community dynamics, and thus characteristics of OSS communties are highly relevant.
