The official announcement of the survey reads:
In my own answers, I have made my best to stress what I believe are two core strengths of the R community as it exists today:
R is a lowly-centralised programming language: it has a list of core developers, known as “R Core“, but it has no “benevolent dictator for life” like other programming languages. Instead, it has someone whom one might want to call a “benevolent contributor for life” in the person of Hadley Wickham, who undoubtedly deserves some kind of lifetime achievement from the R community for developing ggplot2 and the tidyverse (originally nicknamed the “Hadleyverse” by others).
Similarly, R has not one but many package archive networks, including of course CRAN, but also Bioconductor and GitHub, the latter of which brings the virtuous entropy of a place like the Amazonian forest to the R language. As one of my previous note should perhaps have made more obvious, I strongly believe that this ’anarchic’ state of the R ecosystem is essential to its diversity and, in the end, good health. In fact, I believe this holds largely true in any complex system, including political systems.
Decentralisation has its costs: GitHub-hosted packages and alternative engines are places where one might easily inject malware, and diversity means that users get lots of (possibly redundant) choices where they might favour a more restricted set of options. But even those costs have positive externalities in the form of open source software vetting and the development of intelligent safeguards, such as sandboxing and company-monitored programming environments such as Microsoft R.
Another characteristic of the R community that only got a brief mention in a previous note is its humaneness, which I encompasses many qualities, including special attention to the tackling of gender, race, physical and sexual discriminations – to cite only a few – commonly encountered in other social environments.
Just like the R community lacks a (hopefully benevolent) dictator, R-Ladies Global is the kind of initiative that certainly lacks in many programming communities. For a programming language to find its fullest range of locutors and reap the benefits of cognitive diversity, foundations and consortia are not enough: support groups are necessary to enable participation.
Justice is a sufficient condition to support those initiatives, yet here also, there is a strong positive-externalities argument to be made for difference and diversity from the viewpoint of complex systems, with reference to the arguments of people like Scott E. Page or, from a more sociological angle, Rogers Brubaker.
The ideas outlined in the paragraphs are expressed in simplistic form and will very easily lend themselves to criticism. I suspect, however, that a longer discussion of cognition, diversity and efficiency would reach the exact same conclusions, contra engineering-style arguments rooted in cheap calls to eugenics, meritocracy, natural selection, and the so-called “optimisation” of social systems through similar processes.
It has been routinely observed, at conferences or elsewhere, that the R community includes many non-programmers. My hope is that this observation is true, and that it will stay so. Consequently, I hope that the responses to the R survey will ensure that the qualities required to maintain this state of affairs get proper representation in the future objectives and priorities of the R Consortium.
Side note: While writing this note, I was unable to find the name or online presence of the working group that focuses on inclusiveness in the R community, beyond gender diversity. I would appreciate if anyone could help me identify that group, in order to link to it from this note. : the
group task force is called R Forwards. Thanks to Olivia Brode-Roger for finding it.
: shortly after I published this note, Julia Silge released a set of slides that she presented, with co-authors, at useR! 2017. The presentation, titled “Navigating the R Package Universe,” was initially titled “Navigating the R Package Jungle” – which fits well with my arguments above.