A socio-cognitive analysis of online design discussions in an Open Source Software community
Citation: Flore Barcellini, Françoise Détienne, Jean-Marie Burkhardt, Warren Sack (2008) A socio-cognitive analysis of online design discussions in an Open Source Software community. Interacting with Computers (RSS)
Tagged: Computer Science (RSS) open source (RSS), design methodology (RSS), CSCW (RSS), design rationale (RSS), Open Source Software Community (RSS), cognitive ergonomics (RSS), argumentation (RSS), online argumentation (RSS)
This paper synthesizes Barcellini's previous research on the Python community. Its purpose is to outline a theoretical framework and methodology combining structural and content analyses. While the analysis challenges the Bazaar model of open source software development, and provides potential improvements for structural browsing of design discussions and listservs, its hand-construction is too laborious for routine use. The researchers hope that some of this methodology could be automated, for instance in the Conversation Map*.
This paper presents a theoretical framework for those earlier findings, namely focusing on three questions, from three different areas: social dynamics, thematic temporal dynamics, and design dynamics. A quotation-based graph model is used for presenting discussions of two Python listserv discussions; these discussions concerned design decisions called Python Enhancement Proposals (PEPs).
Question: whether "participants with different statuses take part differently in the online discussions and the design process". (The answer, as we know from following Barcellini's earlier work on the Python community, is yes.)
Methodology: Addressed in 4 ways: number of messages, type of quotation structure, depth of quotation, postion in the discussion.
Key Findings: Roles emerge. For instance, the project leader and the person championing a proposal are frequent contributors and their messages receive multiple responses. Only the champion writes syntheses. There is shared power, both among the leadership and with developers; developers participate "even if discussions are framed by the project leader, the administrators, and the champion of PEP".
Thematic Temporal Dynamics (how themes change over time)
Question: are designs and themes important for online design discussions? Are these discussions different from face-to-face exchanges?
Methodology: Addressed by the themes and the temporal order of posts and their quotations.
Key Findings: Initial messages are multi-theme; later messages are more specialized on a single theme. To ensure they are "timely" (according to community standards messages may be posted quickly, with many messages typically received within an hour's time.
Question: Are the collaborative design activities and sequence the same in online, open source software communities, and in face-to-face design?
Methodology: Addressed by a content analysis.
Key Findings: evaluation is key and clarification may be moved elswhere as "off-topic". "Few decisions are taken during the online discussions, but those decisions taken are explicit. "
In addition to what was proposed in earlier work, this paper gives a theoretical framework, and gives names to ideas discussed earlier, such as
- Sequential structure (quoted by one message)
- Branching structure (quoted by multiple messages)
- Closing structure (not quoted)
- Quotation temporality (delay before a message is posted)
- Theme temporality (how long a theme is discussed)
The design view discussed in A study of online discussion in an Open-Source community: reconstructing thematic coherence and argumentation from quotation practices is further described and slightly revised, with definitions for particular activities:
- Group coordination
- Explicit decision
- Other activities (e.g. humor)
Other material not significantly discussed in Barcellini's earlier work
Additionally, these areas are
Types of collaborative design activities
- Generative-evaluation activities - Process of solving and evaluating design problems
- Clarification activities - (or cognitive synchronization) - Constructing a shared representation
- Group management activities
The results section provides further analysis. Results are discussed in detail in section 6, not analyzed here. (Please add this if you can!)
- Sack, W., 2000. Conversation Map: a content-based Usenet newsgroup browser. In: Proceedings of IUI 2000. ACM Press, New York, USA,
Theoretical and practical relevance:
Challenges the bazaar model, at least for the Python project. Seeks automatic analysis to repeat this work on other corpora.
The relevant argumentation background is that "Co-designers accomplish evaluation activities by arguing with each other." These arguments have been characterized as sequences of moves or turns, see
- d’Astous, P., De ́tienne, F., Robillard, P.N., Visser, W., 2001. Quantitative measurements of the influence of participants’ roles during peer review meetings. Empirical Software Engineering 6, 143–159.
- Stempfle, J., Badke-Schaub, P., 2002. Thinking in design teams – an analysis of team communication. Design Studies 23 (5), 473–496.