(Redirected from Filterbot)

The censor, also known as the bad word detector[1], filterbot or simply filter, filters certain words or phrases and alerts the user that what they wrote is against Community Guidelines. As a rigid rule, profanity is a violation of the Terms of Use and Community Guidelines. As such, all parts of the Scratch Website and Scratch Wiki utilize some type of censor. Additionally, posting references to inappropriate topics or links to unmoderated chat sites are also often censored. In addition, some words trigger a "soft report" where the Scratch Team is notified but the comment is not blocked.[2]

Scratch Website

Main article: Mute

On the Scratch Website, if a blacklisted string is encountered in a comment, it is automatically disallowed from being posted onto the website, the user who posts the comment is automatically temporarily muted, and the Scratch Team is notified. Scratchers used a stickied forum topic to report filter bugs, however, Scratchers are now encouraged to use Contact Us to report these.[3] The filter on the Scratch Website is implemented with Cleanspeak.[4] The filter is also stricter for New Scratchers than it is for Scratchers.[5]

Notable Filtered Topics

Document.png This article or section needs additional citations for verification. Its information may not be accurate. Editors can help by adding references. (June 2023)

In addition to inappropriate words or phrases, the censor also filters out a number of other things that are not necessarily inappropriate but nevertheless unsuitable for Scratch. Notably:

  • The word "discord" is censored regardless of context (one can get muted for commenting that contains that word) since it is the name of a service that allows private or unmoderated chatting.
  • The filter also incorrectly filters out some words in non-English languages, because they may have an inappropriate meaning in English.
  • All strings containing 10 to 16[6] numbers one after the other (without letters between them) are replaced with the corresponding amount of X-s because they could be used for sharing phone numbers.
  • The filter originally also blocked words that were themselves appropriate but contained inappropriate words in them, but this has been fixed by adding a whitelist.

Forums

On the Scratch Discussion Forums, if a user attempts to post something that goes against the censor, it will fail to post, but the user will not be muted.[7]

In addition, there is a list of censored words that are replaced with an asterisk. Prior to November 2023, this was the only system for censoring posts. No notification is sent upon posts with censored words.[citation needed] The asterisk was originally intended to be a stop-gap until the censor used on the rest of the website could be used[8], but it still exists on the forums even after the addition of the new system.

Account Creation Page

The error message that shows when attempting to use a censored username

If someone tries to create an account that contains vulgar or censored terms, "Username not allowed" will be shown to the right of the screen.

Scratch Wiki

Censor of the Scratch Wiki

The Scratch Wiki has its own censoring system, using the MediaWiki extension AbuseFilter. It prevents actions from happening if anything contains inappropriate language, and otherwise blocks guideline-breaking actions.

Problems

Scratch 1.x Website

On the Scratch 1.x website, until late 2011, phrases like "skyscraper", "grape", or "push it" were blocked due to an issue known as the Scunthorpe problem, where a censor encounters a false positive and flags a string not intended to be flagged. Since then, a whitelist was added containing many words and phrases that incorrectly trigger the censor.[9]

Scratch 2.0 Website and later

The censor technology sometimes causes problems in languages other than English. Swear words in those languages (unless spelled exactly the same as in English) are not censored.[10] Also, censored words in English may have other meanings or be included in words in other languages.[11]

With the spam filter, there have been problems such as the censor stopping users from posting innocent links, such as project and studio links.[12]

Censor outage on 2022-04-12

Main article: Censor Outage on 2022-04-12

On April 12, 2022, the censor stopped working completely because of issues related to the system,[13] which resulted in a large insurgence of trolls registering accounts with inappropriate usernames and submitting inappropriate comments and Discussion Forums posts. As the result, all comments were closed and hidden from the public and the Discussion Forums were put under Maintenance Mode until the censor returned back to normal.

Other Uses

  • To reduce commercial spam, the censor has been also known to filter out numbers often used by commercial spammers.
    • This part has been stopped due to problems caused.[14]
    • Similarly, the censor filters out words and phrases commonly used by spammers, such as "escort" and "customer service".
  • The censor has been reportedly filtering out ASCII art.[15]
  • On the forums, the censor has sometimes been used for things other than swear words. Prior to a registration list of user-created websites, the domain names of some popular website creating services, such as Weebly, were censored. Alongside, certain shortened links were censored because the shortened link may link to inappropriate sites or pages with malware.
  • On April Fools' Day, the forum censor has been used for some jokes. In 2011, this was primary, but in 2012-13, the entire website had a joke theme (Neigh and Pico, respectively), and the censoring was a strengthening of the joke.[citation needed]
  • Links to sites primarily for private or unmoderated messaging, including Google Docs and Discord, are censored.[16]
  • To reduce unmoderated chat through livestreams, terms such as "livestream" are censored, and YouTube channels/videos containing livestreams and premiers are strictly prohibited, regardless of chat being disabled or not.[17]
  • To prevent inappropriate sites, TinyUrl, Bitly, and other related link shorteners are censored in comments. They are also not allowed in signatures.[18]
  • Names of some Scratch-related browser extensions are filtered out, as mentioning them is against the Community Guidelines.[19]

Controversies

It has been alleged that the Scratch Team had forgotten to censor certain terms, even major vulgar insults. While some minor profanities remain uncensored, they are still censored in the forums, which confused many to think that it was censored. In addition, using blank tags can also help bypass the censors in the forums.[20] Innocent terms, such as "homophone,"[21] were censored due to typos, but were later on uncensored after the Scratch Team had confirmed that these censors were made due to typos.[22] It is an ongoing debate about whether words that are only a bad word if the context is bad should be censored.

References

  1. "Hmm...the bad word detector thinks there is a problem with your comment. Please change it and remember to be respectful." (scratch:js/splash.intl.js)
  2. topic:6494548[dead link]
  3. post:7668203
  4. scratch:parents
  5. Paddle2See. (2016-02-16). "Keep working towards that full Scratcher status - that helps a lot with the filter." post:1795039
  6. projects:726551890/#comments-365386398
  7. File:Cannot Post To Forums Via Filter.png
  8. Paddle2See. (2022-07-26). "It was supposed to just be a temporary thing until the main filter could be integrated into it - but it has never made it to the top of the “to do” list!" post:6486638
  9. Lightnin. (2011-11-03). "I'll add them to the whitelist so they don't set off the filter anymore." ar-post:983156
  10. lisafireball. (2014-04-22). "I can curse and stuff in a different language." topic:36504
  11. Little_sweet_girl. (2014-03-10). "when I am saying about the Vocaloid Song “Fukkireta”, the anime “Kurosh i tsuji” or…" topic:32001
  12. cs96871. (2016-01-14). I go to post the link to the studio and what do I get? “Hmm…the bad word detector thinks there is a problem with your text. Please change it and remember to be respectful.” topic:178341
  13. Paddle2See. (2022-04-12). "There were some issues with the filtering system but they weren't related to any kind of DOS." post:6206238
  14. Paddle2See. (2016-02-22). "I've shut those filters off for now - it should work for you." post:1808094
  15. Mecha11_. (2017-04-20). "Text Sans is classed as inappropriate language[title]" topic:250794
  16. Paddle2See. (2016-10-07). "Please don't do that as it would be considered unmoderated private chatting." post:2237570
  17. lHatr. (2021-03-07). "I don't believe live streams are allowed because one, of the chatting, and two, you can't moderate what could be coming next in the stream." post:4971824
  18. post:4719530[dead link]
  19. ceebee. (2017-11-30). "We want to announce today that we’re shifting our policy related to browser extensions, userscripts, userstyles, and other similar external software. We’ve decided that, for safety reasons, we are no longer going to allow these to be advertised on Scratch." topic:284272
  20. topic:495897
  21. post:5122097
  22. topic:493537
Cookies help us deliver our services. By using our services, you agree to our use of cookies.