Closed Papershine closed 6 years ago
Are you going to keep the source private or public?
If you're keeping it public, someone can just modify it with few lines of code to make it post in chat.
If you're keeping it private, some users may object to using a tool without proper disclosure. Would you be posting something on meta disclosing this?
@Fortunate-MAN It's a hard choice, but public source would be better for me. Firstly, there would not be any instructions on installing and how to install it. Secondly, why would a user want to waste some time to understand the whole codebase and edit it just to chase users? We could license it under GPL to require people to release their edited source. If we see it, we can try contact them. Thirdly, a moderator-only bot, UserStalker, is open source. Also FWIW it seems that the chance of someone doing this just to see some reports about possible voting fraud is very very slim.
What are you going to do to disclose this (if it works and is actively used) to the community? A meta post?
IMO, things like fraud, autoflagging etc are sensitive topics, and disclosure is important otherwise the community could label us as a group of people who flag purely based on a bot.
@Fortunate-MAN If it was included in MOBotics, I think we should ask the opinions of the mods there if they want to open a meta post, but if it was reported to public chat/a web dash, after rigid testing and filter improvements, I would make a meta post and see the response of the community.
https://chat.stackoverflow.com/rooms/167761/conversation/2018-03-sobotics-room-meeting-topic-5
As discussed at https://github.com/SOBotics/SOBotics.github.io/issues/17, some people (including me) now have second thoughts about creating a bot for voting fraud/sockpuppeting, while some think that probably it's still good to go. This topic is to talk about whether the development of Sockite should continue.
Reasons for why Sockite should be continued:
Reasons for why Sockite should not be continued:
Currently, the filters are to be implemented (and can be implemented possibly with the API) can be found here. IIRC, Magisch's query is quite accurate because of a loophole in SEDE, so Sockite might not be able to produce the same amount of accuracy (?).