Page MenuHomePhabricator

Use a third-party blacklist to avoid explicit words from popping up in the captcha
Open, Needs TriagePublic

Description

I've seen explicit words popping up too frequently in our captcha. This should be taken care of. We can use a 3rd party list to replace ours. For example this.

Event Timeline

Niharika created this task.Apr 17 2017, 8:33 PM
Restricted Application added a subscriber: Aklapper. · View Herald TranscriptApr 17 2017, 8:33 PM
Niharika added a subscriber: Reedy.Apr 17 2017, 8:33 PM
Platonides added a subscriber: Platonides.EditedApr 18 2017, 12:28 AM

We could use several blacklists. Adding that list should be easy. Not that many of those words could be generated by our captcha, though.

I am worried however about your experience of «explicit words popping up too frequently». Could note them when you come across them? Maybe there's a pattern.

And of course, adding a blacklist brings out the old issue that the current captcha is completely English-centric (T7309). The Shutterstock list provided above has the bonus of being available in multiple languages.

Well, I've been testing on our Beta cluster, and according to Reedy, Labs has a shorter bad-word list so it might not happen as often on enwiki. Guess we'd have to find out.

Well, I've been testing on our Beta cluster, and according to Reedy, Labs has a shorter bad-word list so it might not happen as often on enwiki. Guess we'd have to find out.

Log onto terbium and have a look at /etc/fancycaptcha/badwords

What qualifies as "third party"? We have Wiktionary, by the way.

Per this discussion:
https://en.wikipedia.org/wiki/Wikipedia:Village_pump_(technical)#Offensive_Captcha

can someone make sure that "nazis" is added, if it isn't already on the list?

Reedy added a comment.May 25 2019, 6:29 PM

Per this discussion:
https://en.wikipedia.org/wiki/Wikipedia:Village_pump_(technical)#Offensive_Captcha

can someone make sure that "nazis" is added, if it isn't already on the list?

It's not - Let me fork this into a separate task that's easier to get actioned by Operations