Thanks for the discussion, folks. I'm curious to hear what others think as well.
Some responses inline. Neal On Thu, Jun 18, 2020 at 9:24 PM Micah Kornfield <emkornfi...@gmail.com> wrote: > sorry for the multiple posts ... I will also note that there is a lot of > debate on this change on the linked thread as well (and I'm not sure the > actual change will happen soon). > > On Thu, Jun 18, 2020 at 9:19 PM Micah Kornfield <emkornfi...@gmail.com> > wrote: > > > FWIW Discussion on git core on naming [1], seems like it might be > > coalescing around "main". > > > > [1] https://lore.kernel.org/git/20200615205722.GG71506@syl.local/ Yes, and there are some reasonable arguments in there for why "main" is a better choice than other alternatives. I was surprised how little bikeshedding there was. > > > > > On Thu, Jun 18, 2020 at 5:27 PM Micah Kornfield <emkornfi...@gmail.com> > > wrote: > > > >> I'm in favor of trying to align on neutral language within the codebase. > >> > >> On branch naming, I think we should wait a little to see if a consensus > >> converges on a new naming convention at least within Git/Github. GitHub is apparently looking into it as well: https://www.bbc.com/news/technology-53050955 > On a > >> technical level, I'm not sure if automated tooling (e.g. crawlers) > outside > >> of the project might make assumptions about default branch names or > what > >> is available in the github API for this type of metadata retrieval. > "default_branch" is already an attribute of "repository" objects in GitHub API responses > >> > >> Thanks, > >> Micah > >> > >> On Thu, Jun 18, 2020 at 1:48 PM Wes McKinney <wesmck...@gmail.com> > wrote: > >> > >>> On Thu, Jun 18, 2020 at 3:33 PM Antoine Pitrou <anto...@python.org> > >>> wrote: > >>> > > >>> > > >>> > Hi, > >>> > > >>> > Le 18/06/2020 à 21:56, Neal Richardson a écrit : > >>> > > Hi all, > >>> > > As you're likely aware, there's growing momentum in the developer > >>> community > >>> > > to drop terminology that some find offensive. > >>> > > >>> > Yes. Is it reasonable? Does it achieve anything? Is there any > sense > >>> > in trying to "drop terminology that some find offensive"? > >>> > >>> We wish to create a community that is open and as inclusive and > >>> welcoming as possible. So yes, IMHO if there is something that some > >>> people might find offensive (even if it is not intended that way), > >>> then there is value in removing that possibility from the equation. > >>> We're here to build a healthy community that builds software together > >>> and so respecting the perspectives of others (even if we disagree with > >>> them) is a part of having a healthy community. > >>> > >>> > > > >>> > As a project that takes pride > >>> > > in being welcoming and inclusive, I think this is something we > >>> should get > >>> > > in front of--particularly as we're approaching a 1.0 release. > >>> > > >>> > I don't think we would get "in front of". We would just be following > >>> > the "growing momentum". In other words, we would do something > because > >>> > it's popular. > >>> > >>> Repeating sentiments from my response a few minutes ago, I think it is > >>> better for us to avoid even the possibility of these concerns arising > >>> in this project. Let us spend our energy debating technical issues > >>> rather than social or political ones. > >>> > >>> > (I'll note that the urge to follow the "growing momentum" is how the > >>> > developer community standardised on irritating tools like Git) > >>> > > >>> > In the long term, and in the face of the problems that it claims to > >>> > address, this seems futile to me. But it makes some people feel good > >>> > about doing something, and it's (small) PR for the project... > >>> > > >>> > Now to the specifics: > >>> > > >>> > > Specifically, I am proposing to: > >>> > > > >>> > > 1. rename the "master" branch to something else ("main" seems to be > >>> > > popular; other version control systems use other words too). > >>> > > >>> > I used Mercurial before Git, and Mercurial uses "default". I used > SVN > >>> > before Mercurial, and SVN uses "trunk". I don't remember if CVS is > >>> > sophisticated enough to have any name for this concept :-) > >>> > > >>> > The problem, though, is that "master" is the overwhelming convention > in > >>> > Git land. Well-known conventions make a better user experience (you > >>> > clone a git repo, you get the "master" branch and you know it: done). > FWIW when you clone (from GitHub at least), you get the default branch, whether it is named "master" or not. > >>> > > >>> > If we choose a non-"master" name, we add an additional hoop to jump > >>> > through for users to approach Arrow. It's a small thing, but > usability > >>> > is often about such small things. > >>> > >>> I'm not concerned about this, given that Arrow is already on the > >>> sophisticated end of the spectrum for open source projects. > >>> > >>> > > 2. replace "whitelist"/"blacklist" in our code with something like > >>> > > "allowlist"/"blocklist", or otherwise renaming. > >>> > > >>> > "allow"/"deny" sounds terser, and also seems more symmetric to me. > >>> > Also, be careful: "block" is very close, unsafely close, to > "black"... > >>> > > >>> > Regards > >>> > > >>> > Antoine. > >>> > >> >
