Improve bad word matching

[DanVanAtta]

Overview: This issue is tracking notes/discussion on improving the bad word matching.

Background discussion:

• https://github.com/triplea-game/triplea/pull/4869#issuecomment-499615008
• https://github.com/triplea-game/triplea/pull/4869#issuecomment-499626919

Code reference:

• https://gist.github.com/Jire/4aa72bd3554cdccdc369c216a230ee56

My 2 cents on a suggested approach:

• begin with a unit test to create a set of bad words and test values that we would expect to match
• create a bean module that implements a matching algorithm and passes the test
• inject the bean module into lobby login validator and use it.

[RoiEXLab]

Reminds me of this: https://youtu.be/CcZdwX4noCE Not sure if it'll be worth investing time in a system that users will find out how to trick eventually 🤷‍♂

[tvleavitt]

Definitely an open question. I understand the impulse to pre-emptively block obnoxious usernames to prevent the system from being flooded, but what do we gain by having a bad words table vs. moderating and blocking individual users, IPs, usernames, what have you? Can usernames be in unicode? If not, we're being rather ethnocentric. Are we going to block every version of "u" as part of the f word? û ü ù ú ū and u? ç ć č and c? I see there are 9 different variations of both "a" and "o" (I get a list by holding down the corresponding key in Gmail). Will moderators know how to reproduce unicode variants for input into the table? I often have to resort to copy and paste when I want to reproduce one myself... will that be easy? I just checked Wikipedia, and there are 137,000 different recognized Unicode characters (out of a possible 1 million plus).

Thomas

On Thu, Jun 6, 2019 at 3:34 PM RoiEX notifications@github.com wrote:

Reminds me of this: https://youtu.be/CcZdwX4noCE Not sure if it'll be worth investing time in a system that users will find out how to trick eventually 🤷‍♂

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/triplea-game/triplea/issues/4871?email_source=notifications&email_token=ABB5HANJDAAW677ZEQ5N6ZTPZGGGPA5CNFSM4HVIAC42YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODXELJQQ#issuecomment-499692738, or mute the thread https://github.com/notifications/unsubscribe-auth/ABB5HANAWG7UMGWPDEVFXG3PZGGGPANCNFSM4HVIAC4Q .

-- Thomas Leavitt Internet enabled since 1990

[stale[bot]]

The TripleA team regrets this issue was not solved sooner. To keep focus on latest issues and the most pressing tasks, this issue has been automatically marked as stale because it has not had recent activity. If the issue may be closed, please do so. If there are remaining items, we encourage for those items to be resubmitted as new, independent, concise tasks. The strategy is divide and conquer. Thank you for your contributions.