Hello! This is just a quick suggestion for Scratch 3.0!
Many scratchers (including me) would probably love to make a cloud chat.
I understand that:
1) It is very difficult, due to cloud data not supporting words and letters, and
2) You must make a ‘white list’; a list of good words (not a bad word list, as this would get you reported) so that the chat is safe.

So, Scratch Team, (and anybody else who finds this) I suggest that you implement the same swear word detector you have on the website to the editor.
And then, make cloud variables support letters and words! This would be so much better for not only safe chats, but also leaderboards and other projects that use cloud data.

-@StinkbombGames

Overall, it's a good idea, but would be very, very tricky to implement, and the downsides outweigh the benefits.

There are reasons that a private messaging system would be bad, aside from swear words. Here are a couple of examples of what could happen:

You are a terrible person. You are the worst person ever. You will never amount to anything.

This is not okay because, even though it doesn't use swear words, one user is still being disrespectful towards another.

Hey, we should talk in person! Here's my email: *email address*.

This is not okay because it shares personal information, which is not allowed on Scratch.

It is insanely difficult to moderate/report disrespectful private messaging, so the Scratch Team has rejected it unless it has a whitelist:

Rejected Suggestions wrote:

1. Allowing free chats with cloud variables/Allowing blacklists:
Allowing “free” chatrooms (i.e. where users can type in whatever they want) are not being implemented because of the potential for bullying/inappropriate messages, and there is no good way to moderate them. Allowing blacklists is also not going to be implemented because it can be way too easy to work around them since there are so many workarounds for a word (for instance, if “cat” were a blocked word, you could work around it with “c@”, “kat”, “katt”, “catt”, “c.at”, “ca.t”, etc.), and because if a user were to look inside the project, they'd see a list of swear words. If you want to make a chatroom, you still can - however, you'll have to make a whitelist; that is, a program that only allows you to select from certain messages.

2. Private messaging: When communication is public, people are more likely to be respectful because they know that everyone can see it. However, when posting PMs, people know that only the intended recipient can see it, so don't think as much before posting. Even if a “Flag PM” function is implemented, the Scratch Team currently does not have the resources to moderate it, because of the reason I said before there would be a lot of inappropriate/disrespectful messages.

A swear word detector for the editor would be impossible, since the workarounds would be really easy. I have a detailed list of such workarounds in this post: https://scratch.mit.edu/discuss/topic/325696/?page=1#post-3360447

I understand that something like that could/may happen, but the same thing could still happen if the chat used a whitelist.
For the email/personal info problem you do have a valid point.

1) See Za-Chary's answer
2) Here: https://scratch.mit.edu/projects/70179488 this is a pretty good wordlist for cloud chat projects, in my opinion.

Za-Chary wrote:

Overall, it's a good idea, but would be very, very tricky to implement, and the downsides outweigh the benefits.

There are reasons that a private messaging system would be bad, aside from swear words. Here are a couple of examples of what could happen:

You are a terrible person. You are the worst person ever. You will never amount to anything.

This is not okay because, even though it doesn't use swear words, one user is still being disrespectful towards another.

Hey, we should talk in person! Here's my email: *email address*.

This is not okay because it shares personal information, which is not allowed on Scratch.

It is insanely difficult to moderate/report disrespectful private messaging, so the Scratch Team has rejected it unless it has a whitelist:

Rejected Suggestions wrote:

1. Allowing free chats with cloud variables/Allowing blacklists:
Allowing “free” chatrooms (i.e. where users can type in whatever they want) are not being implemented because of the potential for bullying/inappropriate messages, and there is no good way to moderate them. Allowing blacklists is also not going to be implemented because it can be way too easy to work around them since there are so many workarounds for a word (for instance, if “cat” were a blocked word, you could work around it with “c@”, “kat”, “katt”, “catt”, “c.at”, “ca.t”, etc.), and because if a user were to look inside the project, they'd see a list of swear words. If you want to make a chatroom, you still can - however, you'll have to make a whitelist; that is, a program that only allows you to select from certain messages.

2. Private messaging: When communication is public, people are more likely to be respectful because they know that everyone can see it. However, when posting PMs, people know that only the intended recipient can see it, so don't think as much before posting. Even if a “Flag PM” function is implemented, the Scratch Team currently does not have the resources to moderate it, because of the reason I said before there would be a lot of inappropriate/disrespectful messages.

A swear word detector for the editor would be impossible, since the workarounds would be really easy. I have a detailed list of such workarounds in this post: https://scratch.mit.edu/discuss/topic/325696/?page=1#post-3360447

You can make bad word detectors without workarounds or a bad word list by using translate blocks. Here’s a link to one of my projects using this https://scratch.mit.edu/projects/584121109/

This has already been rejected, right? And cloud chats are really difficult to moderate anyway.

Fudgecrumb wrote:

Za-Chary wrote:

Overall, it's a good idea, but would be very, very tricky to implement, and the downsides outweigh the benefits.

There are reasons that a private messaging system would be bad, aside from swear words. Here are a couple of examples of what could happen:

You are a terrible person. You are the worst person ever. You will never amount to anything.

This is not okay because, even though it doesn't use swear words, one user is still being disrespectful towards another.

Hey, we should talk in person! Here's my email: *email address*.

This is not okay because it shares personal information, which is not allowed on Scratch.

It is insanely difficult to moderate/report disrespectful private messaging, so the Scratch Team has rejected it unless it has a whitelist:

Rejected Suggestions wrote:

1. Allowing free chats with cloud variables/Allowing blacklists:
Allowing “free” chatrooms (i.e. where users can type in whatever they want) are not being implemented because of the potential for bullying/inappropriate messages, and there is no good way to moderate them. Allowing blacklists is also not going to be implemented because it can be way too easy to work around them since there are so many workarounds for a word (for instance, if “cat” were a blocked word, you could work around it with “c@”, “kat”, “katt”, “catt”, “c.at”, “ca.t”, etc.), and because if a user were to look inside the project, they'd see a list of swear words. If you want to make a chatroom, you still can - however, you'll have to make a whitelist; that is, a program that only allows you to select from certain messages.

2. Private messaging: When communication is public, people are more likely to be respectful because they know that everyone can see it. However, when posting PMs, people know that only the intended recipient can see it, so don't think as much before posting. Even if a “Flag PM” function is implemented, the Scratch Team currently does not have the resources to moderate it, because of the reason I said before there would be a lot of inappropriate/disrespectful messages.

A swear word detector for the editor would be impossible, since the workarounds would be really easy. I have a detailed list of such workarounds in this post: https://scratch.mit.edu/discuss/topic/325696/?page=1#post-3360447

Look, Scratch already made a law of sorts which only allows chat interfaces with options to prevent this.
But, some people will see the cloud data, decode numeric messages, and see private conversations. To prevent this, detectors should be put which use cloud data and all to:
Crash the project
Add the username to some cloud data, so that the creator of the project is notified when someone tries to penetrate the security system.
I have implemented this strategy in the following project:
https://scratch.mit.edu/projects/1097068123