search

glam-lab / degender-the-web

A Chrome extension that replaces all gendered pronouns with "they/them/their."
GNU Affero General Public License v3.0
6 stars 3 forks source link

It's blocked on too many pages that happen to have the word "gender" #117

Open beszel opened 4 years ago

beszel commented 4 years ago

Is your feature request related to a problem? Please describe. DGtW is very conservative about which pages it runs on in order to avoid replacing important pronouns (#34). This causes it to be blocked on many pages that do not discuss gender, with no option for the user to re-enable it.

Describe the solution you'd like The word "gender" should act only as a stopword in the title of a page. "Gendered" and "genderqueer" should still act as stopwords, as they're more likely to show up on pages that discuss gender.

Describe alternatives you've considered The options could instead give the user a whitelist of pages and a whitelist of domains, with two buttons in the popup to quickly enable the extension on the current page or domain. Use of these buttons would save the current page or domain to its respective whitelist.

ProfJanetDavis commented 4 years ago

The proposed solution would mishandle two of the three test cases proposed in #34. I like the whitelist idea, but would need to think it through.

Here are some more alternatives:

  1. Search for "gender" as part of longer/compound words (gendered, transgender, genderqueer) but not as a complete word. That's technically straightforward to implement with a regex, but would leave out phrases like "gender transition."
  2. Search for a list of keywords/phrases including "gender". We'd have to come up with a good list. We might want to do that anyway to evaluate whether (1) is a good solution.
  3. Require that "gender" appear on the page more than once. I've thought about this before. It would account for the "Terry Pratchett" linked above.
  4. Use some more sophisticated NLP technique for topic analysis to determine if gender is a significant topic of the page.

Can you think of others?

Can we start cataloging pages here that include the word "gender" but aren't really about gender?

beszel commented 4 years ago

Alternative 1 is what I meant to propose; I like the look of 1 and 2 best, and I think we'll learn more about the others once we find more pages like this. I'll keep an eye out for some examples.