Gendered Language: Feature or Bug in Software Documentation?

[ericdill]

I was inspired by https://modelviewculture.com/pieces/gendered-language-feature-or-bug-in-software-documentation.

This PR adds tests to all docstrings in our library to prevent gendered language from leaking in

[ericdill]

Also inspired by this https://modelviewculture.com/pieces/the-open-source-identity-crisis

[danielballan]

For a practical review of ways to avoid gendered assumptions without resorting to contorted English, see Brian Garder, author of "Modern American Usage" in a series of three short posts. The most relevant is this one. @ericdill maybe add a link to this PR?

[danielballan]

Is full-on blacklisting the male pronoun is the right way to ensure openness in the library? Sometimes we might want to refer to the work of a specific researcher who has a gender. Maybe a warning would be better than a test failure.

[tacaswell]

👍 on this.

I am inclined to say we deal with the case where we must use a pronoun to refer to a researcher in a docstring at a when it comes up. I suspect that if we have enough text in the docstring that using a pronoun is required there is too much text and it should move off to the rst documents anyway.

[danielballan]

It looks like there is some accidentally-committed code in here.

[ericdill]

It is not accidentally committed. This new test actually imports every module and apparently some of our relative imports were incorrect

[ericdill]

@danielballan Added a link to the writing guidelines. Thanks for that

[ericdill]

Also, this should really make use of regex, but I suck at it