The Problem and Solution
The current filterbot has one system for all the bad words and phrases to protect users. However, not every bad phrase is the same offensiveness, and some of them can be used accidentally, which is a problem if you're getting muted for 5 minutes for each of them. This suggestion is suggesting that the one filterbot is replaced with multiple, for each different kind of bad phrase.

The Important Filters
However they choose to implement this, it should have at least two kinds: A minor and major level of filterbot.

• The major level filterbot is similar to the one we have now, muting you. These are for your “definitely trying to offend someone” phrases that can't be used accidentally, such as slurs.
  
• The minor level filterbot only prevents the comment from being posted without muting the user. This is for things like private messaging apps or browser extensions, because not many Scratchers know these aren't allowed.
  

The Other Filters
Those were the basic two, but I have two more ideas that could also be implemented. These don't have to be implemented if there are problems with them, as long as the main two are implemented.

• The warning level filterbot warns the user first, and if they try to post the comment again, they get muted. These are for stuff like “middle finger”, which can be used accidentally (I used it in a reply to someone explaining something without knowing it was offensive). They might not be aware the first time, so they get warned and think “oh, there's something offensive in my comment, let me revise that”. If they insist, they get muted.
  
• The review queue level filterbot sends the comment to a review queue that the creator can approve or delete the comments. Most of the very offensive stuff are blocked by the major and minor level filterbots, so the remaining comments probably won't be too bad.
  

Misconceptions

• This suggestion isn't a choice between multiple filterbots. This suggestion is a suggestion that the Scratch Team changes the system of how each phrase is treated when you comment, so that instead of everything muting you, some phrases give mutes, some phrases just stop the comment from being posted, and some phrases warn you before muting you.
  

I think the major filterbot should only apply to words and phrases that are 100% certain, i.e. no spaces or other characters between them. If it catches anything else it goes to the minor one.

Bump

Support, but I think there should be warning before being muted by filterbot, I recently got a days muting by copy and pasting some wordart that I had posted years earlier, when it was allowed.

Techno-CAT wrote:

Support, but I think there should be warning before being muted by filterbot, I recently got a days muting by copy and pasting some wordart that I had posted years earlier, when it was allowed.

hi dark mode suggestion guy
I think stuff like that would be in the minor filterbot, but the warning would be good. But I think there should be 3 levels for that. I'll change the siggestion.

Bont

Bump

Yes. This is a great idea, easy set-up and definitely a soulution to a problem.

I've heard a rumor that some words like n word is just an appropriate expression in other languages and these people using other languages just got banned because the systems they use think it's n word

week bump

4 day bump

bum p

Bump

LittleOctopusIsATest wrote:

I've heard a rumor that some words like n word is just an appropriate expression in other languages and these people using other languages just got banned because the systems they use think it's n word

How would it know otherwise?

LittleOctopusIsATest wrote:

I've heard a rumor that some words like n word is just an appropriate expression in other languages and these people using other languages just got banned because the systems they use think it's n word

So it would go in the major level with warning.

Support! This would cut down a lot on the "help i got muted/paused for saying [word]" questions and bugs and glitches topics.

b'mup

PkmnQ wrote:

There should be at least 2 levels of filterbot.

You could also have a review queue level where the creator could determine if the comments are ok or not. There are lots of possibilities for this.

For this, I no support, since then people could see the not okay comments.
However, if you mean the ST, then I do support, as that is safe.

legobuilder000 wrote:

PkmnQ wrote:

You could also have a review queue level where the creator could determine if the comments are ok or not. There are lots of possibilities for this.

For this, I no support, since then people could see the not okay comments.

Only the creator of the project can see them, and most of the not okay comments would be caught by the major level of the filterbot.