Dalle, David et al, 2004
It took me a while to realise it, but this paper is sketching the research agenda: it contains few findings as such. However, it is the best overview of research approaches that I have found so far and I will be following up on citations to and from it and its authors.
In this paper the interaction with commercial sponsors is touched on as a separate research area, if only briefly. The emphasis is on the motivations of the individual developers; it seems that because SourceForge and the like make it easy to analyse developers' work (eg J Mateus Garcia1), that's where the research is directed.
Introduction and context
The paper starts with the economic issues: the motivation for developers to do unremunerated work and the secret behind F/LOSS projects’ ability to compete with proprietary software
It asserts that "F/LOSS movements carry broader economic and social significance and therefore deserve to be the subject of continuing… study". FLOSS is placed in the context of "invisible colleges" of the 17th century – it the scale and dispersion of the participants that is unprecedented, not the basic concept*
The OS research programme is operating in the context of a movement of information goods to centre-stage and a wider use of peer-to-peer modes of working,** increasing the benefits of the open conduct of work over secretive practises.
Finally, there’s the "abundance and accessibility" of data on the development process§. Again, build the data, and they will come…
Measurement of the utility of F/LOSS programs is seen as the starting point: driven by developers' perceptions of what they're working at. One has to ask: why should we go for a development process so skewed to the desires of socially inadequate young men?§§.
Main elements of the inter-relationships between of the elements of a F/LOSS project have been sketched out, for example in the research approach by Stanford, supported by Paris VI, Maastricht and Sussex Universities in Europe.
Dividing the research field
The research field can be divided into four Organisational Economics aspects:
Patterns of code committing
Testing power law assumption (many commits by few developers), examining clusters of developers. Focussed on Linux kernel, even though it's admitted to be atypical
The match between developers' motivations and final users' needs
Uses SourceForge stats to analyse what makes a project successful – balance of lock-in vs developer turnover. SF data can be filtered by tool and technology. Projects are categorised between I-mode and C-mode – most are I-mode, as would be expected
Relations between developers and sponsors
Quotes FLOSS-US3 as most recent source of information on finance and motivation of developers. Ghosh was involved in the original FLOSS report2
The final area of proposed research is computer simulation the effects of varying reward factors: eg choice of kernal vs application project etc. The aim is to make the model itself OSS, and improve it with input from other research fields.
Future directions of research [pp17-20]
The authors see OSS as a paradigm shift comparable to the creation of shrink-wrapped software in the 80s: it is a communal ethos for creating 'public goods'. My immediate thought, that public goods are creating private income for free-riders (eg IBM or any other consultancy), is addressed with the observation that profit-seekers appear to be willing to subsidise the production and distribution of OSS.
It is pointed out that science, another universalistic/distributed activity, is dependent on patronage or a symbiotic relationship with funders.
In parallel, the Stallman-inspired "perfectionist impulse" is seen as working with a demand-pull "instrumental need" by the market for a stable network infrastructure. OSS's future will be determined by the balance of the inevitable weakening of the former and the possibly growing pull of the latter. Success is not seen as inevitable; the authors also allow for a kind of glorious failure, whereby the OSS movement dies, but only after permanently changing the industry.
Factors in future developments of OSS:
- How XML develops as an information-exchange language. It may end up as a series of proprietary standards, in which case, OSS will become a ‘low-level’/infrastructural activity only
- Impact of peer-to-peer networking on IP rights and value creation/destruction
- How important status rather than remuneration-oriented motivations will remain. Is the tribal-membership aspect important for people who would otherwise be isolated in their work and professional environment?
- Impact of OS distribution on concepts of publication in general
- J Mateus Garcia. A protocol for the Web Mining of Sourceforge, 2004. SPRU, University of Sussex
- FLOSS, Free/Libre Open Source Software: Survey and Study. 2002, International Institute of Infonomics, University of Maastricht: Berlin.
- David, P.A., A. Waterman, and S. Arora, The Free/Libre Open Source Software Survey for 2003. 2003, Stanford University: Stanford, CA.
* It is interesting to note that the members of the invisible colleges were largely landed gentry or other men of independent means. Some aspects of programming could be argued to be the domain of the gifted amateur; so maybe there is an analogy with relationship between science (programmers) and industry (consultants)?
** Though I would have thought the invisible colleges and other scientific networks were also peer-to-peer. The packets that were exchanged were physical, not electronic.
§ From the point of view of the developer, not the organisation funding the process
§§ Of which I was one. Now I am merely no longer young