30 Oct 2000 renster   » (Journeyer)

When I become an apprentice I'll post this here as a question to developrs

I'm currently starting a PhD at an Australian University and I want to examine the open source model of software development. I am interested in using strategic management theory (using a resource-based view) to identify factors that lead to successful open source software projects. After doing some reading, and discussion with my supervisors, I am very keen to move my research along. I do in fact have lots of questions relating to research on the open source model that I have to address in the near future. I would like to ask the Slashdot community to comment on one issue in particular and it is a fairly fundamental question.

I've set up my phd as a project on sourceforge (phdstrat.sourceforge.net). I'll be using the site as my main information publishing venue etc. I have also put up my draft research outline (http://phdstrat.sourceforge.net/stage2.html) which I would also like any comments on. It is far from being finished or finalised.

My main question at the moment is, what is `success' when looking at open source software development projects? How can you tell one project is more effective than another? This is a critical question for researchers who wish to examine open source software projects. I haven't really seen the issue addressed in any detail and have it going over in my mind. In particular it is an issue for the management field as most researchers examine firms where success is measured in dollar terms (profit, ROI etc.). This is rather a narrow view of success (effectiveness or performance) in itself and won't be appropriate for examining open source software project success - for obvious reasons.

In the recent open letter by Bob Young to Slashdot, Bob mentions Redhat's success is derived from the open source license which gives customers control over the technology they are using. There are a range of other properties of the open source model that have been pointed to as reasons for its success. These are often made by comparing open source to the proprietary binary-only model. As with Rob Young I am not interested in this comparison at this point in time as much more work needs to be done on understanding the open source model itself. Yes, I think it is interesting and necessary, but I want to focus on what distinguishes a successful or effective open source project from an unsuccessful or ineffective open source project. I'll leave the detailed open source - close sourced comparisons to someone else.

I think that the open source model requires that we develop a broader definition of `success'. Exactly what that definition is I'm not sure. The documents I have read point to a range of indicators of effectiveness or success and point to Apache, Linux etc. A resource-based approach assumes that project leaders will attempt to make their project successful. Success is contingent on the strategic choices they make in relation to the identification, development and deployment of key resources. Given this, projects should vary in terms of the set of resources they have. Some projects will be more effective and more successful; others will be less effective and less successful. There will be many resources, however only some resources will be "key" resources. As there has been no research on open source projects using a resource-based view exactly what the resources are is difficult to define at this point in time - a blank canvas if you like.

We can maybe look at the benefits to the individual to see what might contribute to project success. I can definitely see pay-offs to individuals in terms of developing knowledge and skills, so is effectiveness based on getting the most knowledgeable and skilled people on your project? Successful projects might be ones where project leaders are able to draw in talented programmers because they are themselves talented and respected programmers. Perhaps projects are successful when they attract more human, social and organizational resources in general.

As ESR mentions there are also network effects and network externalities operating in the open source model, so is success the size of the network and the amount of resources the network attracts? Is success indicated by network activity? This would indicate some idea of a critical mass being required for a project to begin operating effectively. Naturally this will depend on the type of project being examined (this is another issue I have to deal with).

At the project level project success might be more difficult to define. Does a measure of effectiveness include indicators of how quickly bug reports are resolved, age of the project, how often new versions are released or meeting project deadlines?

On the flipside, what are the most likely causes of project failure or indicated ineffective projects?

I am interested in reading some discussion on this topic. The open source model challenges many assumptions held in the management literature. I think strategic management might have something offer in better understanding open source projects. In the near future I hope to develop my research topic further and even further down the track try to get more information from projects and involve them in my research.

Latest blog entries     Older blog entries

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!