Throughout the years, the community has displayed their dissatisfaction with the ROBLOX filter countless times. To my opinion, adding a filter to the Devforum seems like an awful idea.
Sure the quality of the forum continues to deteriorate but I would rather let that happen than having to deal with a stricter filter on this forum.
Discourse’s filter is nowhere near Roblox’s filter. Devrel has already configured the filter but it needs more to block
Since when has the forum deteriorated itself? Sure, #development-discussion isn’t too hot right now, but it doesn’t mean the whole forum is falling a part.
also if outright rejecting posts, requiring their approval, or blocking out swear words that might be acceptable or might not be; the system can automatically flag posts if they trigger a certain filter.
roblox’s filter is actually really good and uses powerful cutting-edge technology. it’s just people are upset that the filter… does its job.
Support.
I would also like to add on when certain media (such as images) are being uploaded to the CloudFront, such media should go through filtering like OCR to well, prevent this:
With the DevForum being increasingly more easy to gain access to, putting measures like these in place can ensure a professional and non-toxic environment, which I assume is Roblox’s primary goal for this.
The system actually relies on artificial intelligence, which will never be perfect. As it’s not perfect, it’s very easy to pick on it for minor issues, in the long run, they’ll probably be fixed but it’s a bit pointless to just throw comments like this at a system as dynamic as this one.
It’s deliberately strict so Roblox can keep compliant with certain child-related measures to prevent certain restrictions from being placed against Roblox.
Support on this.
Also, I am wondering can this filter out callouts? ( just saying because callouts are getting worst now )
Please keep in mind this user has been banned from the forums. It should have been common sense that, if he did that he would be banned. However for those wondering what happened to the user. Apologies it wasn’t caught faster.
At least the user was punished. Will there be any measures put in place to prevent things like that from happening in the future?
There is measures to prevent this already, however it’s a computer program not perfect against a person trying to use all the possible loopholes as mentioned before by PeZsmistic
We shouldn’t not add this because a few users might bypass it, any kind of protection helps.
I think it would be better to make a smarter decision than slap a filter on the problem when we know it won’t stop anything. There is no difference between a user seeing SLUR and SL_UR, so there is no great pressure to add a bunch of slurs to the blacklist right now when they can be bypassed in many thousands of ways; nothing will be prevented.
Making it well known that certain bad words are filtered just makes people try to bypass them immediately, when instead we could be leveraging their willingness to just try posting them. It would be more effective to leave the forum the way it is, and make plans to build a plugin that that leverages use of certain extreme language to immediately lock down the accounts posting for moderator audit.
Could one not argue then that this is a problem you should stop at the roots by increasing entry requirements to the forum? Im pretty sure there is an obvious correlation between reported instances of this happening and reported counts of members joining the forum, perhaps some sort of individual screening of everyone who wished to join the forum would be a better solution.
Support.
This is rlly getting annoying seeing posts like this in the devfourms.
For example as the tweet shows
Under what specific circumstance would a racial slur be required to have free and open discourse on a forum mostly about game development for a platform that you could normally be banned for saying said slurs anyways?
No idea what you’re referencing to but let me just inform you that I despise people who use swearwords based on race.
What I am however advocating against though is a stricter filter being imposed on this forum. A filter similar to the one we have in the Roblox website.
Nobody is saying a filter like the Roblox website. but we need more filteirng.
The problem at hand was the lack of a decent filter for the devforum(or really one at all), which was resulting in the previously shown screenshots. A powerful filter like the one on the website would surely be a dumb idea to add here since almost all of us are teens and should be capable of not getting upset over certain words or willingly giving out private info, but posts that are intentionally inflammatory and that try to anger people by abusing the lack of a filter, I.E:(what i was speaking in reference to ->) https://media.discordapp.net/attachments/670372148333772840/769959358791942185/unknown.png?width=1010&height=677 as posted by @pythonsbyte
The primary suggestion for a filter on this thread so far has mostly just been to flag posts for review, should they contain offensive material, not to outright bar people from say, posting a discord link, which might be a problem depending on the context.
In my opinion, certain key-words like racial slurs should immediately put the account that’s posting it under review, I really think there would be little to no problem with that, from what I see almost all of the devs here are well behaved.
The only threat I see is roblox making it like the main site filter, “######” everything out or require that the post be “reviewed by a team of moderators” before its published, but then again this is the developer forums, would they really do that to us?
All of that aside though, I really wonder what the person’s intentions were by posting those things, what their overall goal was because in my POV they gained nothing but wasting their life away trying to troll the forums, if their goal was to get attention, well I guess they achieved that, considering this topic is up for discussion.
Edit: Upon reading the discussion over again, let me clarify what I said earlier.
“In my opinion, words like racial slurs should immediately put the account that’s posting it under review”
I was meaning to say certain key-words, not filter every word, example: nobody here should be saying the n-word at all.
Edit #2: I just realized people could try to bypass the filter by using t̸h̸e̶s̸e̵ ̵t̸y̸p̷e̶s̶ ̵o̴f̴ ̸w̶o̸r̴d̵ ̷g̵e̶n̸e̵r̴a̸t̶o̵r̶s̸
Thinking it over, I really don’t think there are ways to combat this other than flagging the post as always.
Alternatively, we could just raise the requirements on posting, but then that’s not a really good idea for new-comer devs that actually want to post on the forums.
Anyone else have ideas?
I’m not sure there’s a correlation, at least not the extent you’re implying. I think most of the users spamming inappropriate content were compromised. Strict entry requirements won’t solve this.
Someone could probably create bots to reach the TL1 requirements and spam inappropriate content, making the requirements stricter would just delay it. The only way to stop this entirely would be manually reviewed applications (infeasible).
I think there would be a better ROI (for time invested to get a reasonable outcome, with the least amount of manual work) investing resources into stateful filtering (example: users get a score based on past forum activity, lower it is more strict the filter is and likely to filter phrases/sentences which could be inappropriate), whether it’s implemented by Discourse or a plugin created in-house
Above wouldn’t solve the issue of users getting compromised, but this is dealt with fairly quick by DET anyway.
I want to add something here.
Does this blacklisting such as the n-word account for different languages? (I think it does) For example d-word in German means fat, and I don’t believe it’s right to censor or take any actions against users across different languages or users who accidentally wrote in their native language, before being told to move to the forum for their language. I feel that inappropriate words can be knotted into other words as well.
For example, I may say medic and accidentally add a k, and now all of sudden I’m in trouble. Or maybe my spelling or spacing was off and it sounds much worse.
In addition, I feel the system should also account for spaces and no spaces, where the word is used 2 or more times. For example f-wordf-word. I may emphasise an intense word. But the system should be able catch intense words spammed.
But as others have mentioned, there are so many loopholes.
Also I don’t appreciate members being linked to every issue on the forum. I feel anyone can get their account compromised and, anyone can spam inappropriate words because let’s be honest, somebody who goes out of the way to do so, would’ve clearly broken and ignored the forum rules because any sensible person can draw that conclusion. Plus getting your account compromised isn’t you doing it.
