> This doesn't look like it's doing any actual NLP. This just seems to be looking for a "conjunction" or a "prepositions" and replaces text.
Although doing actual NLP would be better (though difficult in JS), using this method of looking at conjunctions and prepositions actually works correctly almost all of the time. However, I am also working on a Python implementation that would work even better by using POS tagging.
> The source code also needs a lot of work.
Yes! Please contibute and help me improve this project :-)
No, as this project essentially runs a client-side search and replace and the results are only visible to the person running the user-agent.
It hinges entirely on the behavior of the person using this tool to use its output in a way others find offensive. A benevolent use of this tool would be to batch-replace gendered terms in documentation [1].
Just like race or species can be deeply personal and an important part of one's identity, so is gender. However, I think we can agree that we don't need/want separate pronouns to identify our and others' race. So why gender?
It is true that pronouns can be deeply personal, and that is one reason you might want to make them gender neutral.
This proces does not strip anyone of (part of) their identity; it does liberate people who do not fall into the gender binary, or who think that gender--much like race--should not always have to be specified. People can of course still be ladies and gentlemen if that is how they wish to identify.
This would have been a great application for something like nlp_compromise or some other NLP library.
The source code also needs a lot of work. If you run that through a linter it'll throw you against a wall.
That dev needs to watch "Javascript the Good Parts"