Academia vs. FOSS: The Good, The Bad, and the Ugly
Mel Chua has been pushing forward on the theme of FOSS culture in academia, and has gotten a lot of wonderful comments, many about why it’s not so simple to just port one culture over to the other. I want to try to compile items from Mel, comments on that post, and a few other sources. The question is: what are the salient differences between FOSS and academia?
I will proceed using the now-standard Spaghetti Western classification schema.
- Universities tend to be more proactive about identifying and aiding newcomers that are struggling, as opposed to many FOSS projects that have high failure-and-dropout rates due to poorly designed scaffolding.
- Academia is much more demographically inclusive. FOSS communities are notoriously imbalanced in terms of gender and race.
- The academic fear of having ones results scooped or stolen results in redundant, secrecy, and lonely effort. FOSS communities get around this by having good systems for attribution of incremental progress.
- Despite scientific ideals, academic scientific research is getting less reproducible, and therefore less robust, because of closed code and data. FOSS work is often more reproducible (though not if its poorly documented).
- Closed access academic journals hold many disciplines hostage by holding a monopoly on prestige. This is changing with the push for open access research, but this is still a significant issue. FOSS communities may care about community prestige, but often that prestige comes from community helpfulness or stake in a project. If metrics are used, they are often implicit ones extractable from the code repository itself, like Ohloh. Altmetrics are a solution to this problem.
- In both FOSS and academia, a community of collaborators needs to form around shared interests and skills. But FOSS has come to exemplify the power of the distributed collaboration towards pragmatic goals. One is judged more by ones contributions than by ones academic pedigree, which means that FOSS does not have as much institutional gatekeeping.
- Tenure committees look at papers published, not software developed. So there is little incentive for making robust software as part of the research process, however much that might allow reproducibility and encourage collaboration.
- Since academics are often focused on “the frontier”, they don’t pay much attention to “building blocks”. Academic research culture tends to encourage this because it’s a race for discovery. FOSS regards care of the building blocks as a virtue and rewards the effort with stronger communities built on top of those blocks.
- One reason for the difference between academia and FOSS is bandwidth. Since publications have page limits and are also the main means of academic communication, one wants to dedicate as much space as possible to juicy results at the expense of process documentation that would aid reproducibility. Since FOSS developed using digital communication tools with fewer constraints, it doesn’t have this problem. But academia doesn’t yet value contributions to this amorphous digital wealth of knowledge.
Have I left anything out?