Open bernhold opened 1 year ago
So, I've spent a bit of time playing with this tool. You can find an example of its use here. There it is erroring on words like master
, slave
, grandfathered
and warning on blacklist
. This is all configurable from a .github/in-solidarity.yml
configuration file which permits us to establish per-term settings for
Seems like its pretty useful. We could, for example, set slave
to error mode of failure but master
to error mode of warn.
It works as a GitHub application (aka a bot). So, its a bit different from an Action which is recruited only at time the associated actions are occurring on the repo. You install it to the repo like you do Travis CI or Azure, etc.
I can proceed with more of this on our actual repo. Should I?
It does seem like these checks can miss context too. It almost seems like we need this integrated with something like a grammar checker. OTOH, maybe the recent plethora of on-line, text AI tools could actually obsolete these more primitive tools for something more robust.
For example, grandfathered
(not the ed
at the end) I think we can all agree is probably only used when using the word grandfather
as a verb. In all cases I think that connects back to its use as a form of voter disenfranchisement and we would want to fail
that use. But, grandfather
by itself may or may not be problematic depending on its use. So, for example...
Without integration with a grammar checker, we can't know how grandfather
is being used in a sentence (a verb or a noun).
After chating with ChatGPT, here are some other potential options
Nothing found that combines grammar analysis with inclusive language to ensure robust rules can be defined.
Nothing found that supports per-file ignore options...which I think we've concluded we've needed in other contexts so probably need here as well.
@rinkug given our change in resources and given that whatever action we might decide take here requires some discussion with the @betterscientificsoftware/bssw-editorial-board, I don't think I can move forward on this any time soon.
This app (not action) might be interesting to help vet the language of submissions: https://github.com/marketplace/in-solidarity
I heard about it in a talk about how the UK Turing Institute uses GH as a collaboration platform for their The Turing Way online book.