|
|
| Exploratory filter; looks for new accounts containing strings with NINE consecutive consonants. Just intended for logging at the moment, to see how common these names are, and to what extent it is a good detector of bad usernames. | | Exploratory filter; looks for new accounts containing strings with NINE consecutive consonants. Just intended for logging at the moment, to see how common these names are, and to what extent it is a good detector of bad usernames. |
|
| |
|
| So far, it's looking pretty good, even at 8 characters. Upgrading threshold to 9 to make it even more selective.-- The Anome, 12 May 2018 | | So far, it's looking pretty good, even at 8 characters. Upgrading threshold to 9 to make it even more selective.-- The Anome, 12 May 2018 |
|
| |
|
| More investigation: now looks at only first edits for these accounts. -- The Anome, 18 May 2018 | | More investigation: now looks at only first edits for these accounts. -- The Anome, 18 May 2018 |
|
| |
|
| Performance improvements -- checking user_editcount will cancel out more actions in this case, so it should come first; similarly reordering other conditions to cancel out more actions; user_editcount < 2 will also be true for blank or false values, so removing !user_editcount to save a condition, and using contains_any for checking multiple actions to save a condition ~MA | | Performance improvements -- checking user_editcount will cancel out more actions in this case, so it should come first; similarly reordering other conditions to cancel out more actions; user_editcount < 2 will also be true for blank or false values, so removing !user_editcount to save a condition, and using contains_any for checking multiple actions to save a condition ~MA |
|
| |
|
| Removed all the filtering on event type and edit count. This is fascinating: the lack of output from the previous suggests that people are creating these random-typing accounts, but never using them... -- The Anome, 19 May 2018 | | Removed all the filtering on event type and edit count. This is fascinating: the lack of output from the previous suggests that people are creating these random-typing accounts, but never using them... -- The Anome, 19 May 2018 |
|
| |
|
| Adding a dependency on wikitext, to see if that lets this filter see edits. | | Adding a dependency on wikitext, to see if that lets this filter see edits. |
|
| |
|
| Adding test to exclude highly repetitive usernames which would already have been caught by the repetition filter -- The Anome, 13 June 2018 | | Adding test to exclude highly repetitive usernames which would already have been caught by the repetition filter -- The Anome, 13 June 2018 |
|
| |
|
| Added a few running-finger-down-the-keyboard patterns. -- The Anome, 25 October 2018 | | Added a few running-finger-down-the-keyboard patterns. -- The Anome, 25 October 2018 |
| Fixed implementation error which had the reverse effect of that intended above -- The Anome, 31 October 2018 | | Fixed implementation error which had the reverse effect of that intended above -- The Anome, 31 October 2018 |
|
| |
|
| Added alphabetical-progression patterns. -- The Anome, 2 November 2018 | | Added alphabetical-progression patterns. -- The Anome, 2 November 2018 |
|
| |
|
| Backed off alphabet/keyboard progression patterns, following feedback. -- The Anome, 3 November 2018 | | Backed off alphabet/keyboard progression patterns, following feedback. -- The Anome, 3 November 2018 |
|
| |
|
| Removed filter that prevented overlap with repetition filter: won't actually increase blocking, just more reporting will be evident here. -- The Anome, 7 December 2018 | | Removed filter that prevented overlap with repetition filter: won't actually increase blocking, just more reporting will be evident here. -- The Anome, 7 December 2018 |
| | |
| | Fix warn/disallow messages; disallow was from filter 887 --SoY 4 Mar 2019 |