Clarification on code flagged for safety review

Updates Announcements
#23

This “Clarification” confuses me more than it clarifies things for me.

Does this apply only to open-source code (i.e. Free Models that are given to the public) or all code in general? If this applies to comments I wrote in code nobody will ever see except for myself, then I’m kind of screwed.

Expletives, offsite links (which are absolutely essential in many cases if I am linking a Stack Overflow question, or some tutorial I followed to write a block of code), and some personal things are all present in code that I know nobody will ever see (not even exploiters, since they only get compiled bytecode). There is no good reason why this should be moderated or even looked at by Roblox, so I really hope that this is not what this thread is referring to.

62 Likes
#24

As others have mentioned, I don’t see a point to this, since we’re the only ones who can see our code.

The only time I could see this being necessary is via Team Create. If Team Create isn’t on, this seems unnecessary. Especially with the prospect of people not trusting this “specially-trained team”, even if they are Roblox employees.

And, as much as we appreciate the transparency, I am concerned that we weren’t alerted of this sooner if this system has been in place for several months.

57 Likes
#25

I am also wondering about this! I have linked my Github page as well as a StackOverflow before (for documentation, as well as background on something). Will this result in me getting a moderation?

15 Likes
#26

Why exactly did it take so long for this to be communicated to us? Shouldn’t we have been notified before the policy was put into place, so we would have the opportunity to clean up our code on our own?

So does this mean swearing is allowed, considering our code is private much like the private sections of the devforums?

75 Likes
#27

I understand where this can be effective, since I have seen a free model once containing a script inside with links to NSFW content. BUT I see it as unnecessary to check code for, let’s say profanity, that only I am going to ever read (non team-create places owned by me)
Also, in my opinion, developers should be notified whenever their code has been manually reviewed, even if the code turns out to have inappropriate content

39 Likes
#28

I’m a little confused. Does this apply to code comments? Do the moderation messages you recieve specify what the content is and where it is in your code?

19 Likes
#29

I am sure that their system could figure out if the code is obfuscated - although it most likely will function like the chat filter system… flagging everything

You did make a really great point about the keys… hopefully they do something about that because I’m not for putting trust in some dude I don’t even know

8 Likes
#30

This is stupid. Some of our scripts are going to have personal/private information because we need it there (API keys, etc.) and these are things we don’t want anyone, including staff, to know about. What people don’t see shouldn’t be moderated.

72 Likes
#31

Has this already been implemented, and if so does anyone know when it was implemented?

3 Likes
#32

I do not approve of this change at all. First of all, why should we trust Roblox mods, who now have been confirmed to be able to read our raw source code, not to anonymously toss information obtained through these powers to… let’s just say… less savory peoples in the community for personal gain?

There is no need to review code for anything as if a script is doing something inappropriate it would be apparent ingame from, at worst, a non-moderator account’s perspective.

Your chat filter and other text based filters are NOTORIOUS for false positives on anything and everything, and your human mod review teams are far from infallible too, so even if the above two points are addressed there are still massive problems with this change.

Also, by confirming this you have also confirmed that you likely have full granular edit access to all our content, which is alarming to say the least. What if a developer makes a change in their game via flipping a switch in a script that Roblox does not like economically, politically, or for some other reason? Are they just going to dispatch a mod to flip that switch back? What if the developer, outside of Roblox, does something that Roblox disagrees with? Will they now trash our scripts?

Congratulations, Roblox. After a long streak of neutral or positive updates (Massive props on all of those, by the way) you’ve again given me reason to distrust the platform and reconsider my participation in continued Roblox development.

Edit:
Oh, and by the way, I noticed that you used the actual Roblox account to post this, like you have on pretty much every controversial and negative change to the platform. So far, this gives me reason to believe that everyone else’s complaints regarding this issue are pointless and that you have zero intent on doing ANYTHING in response to community feedback.

Nice job, Roblox. I thought the increasing number of polls meant you were listening a little better to your community.

170 Likes
#33

Yes. They are just now publicizing that is has been implemented. Cant find the DevForum post on it but if I recall correctly there was a game that got moderated for having its own chat filter (script was disabled too)?

I might be remembering wrong, if anyone knows I’d like to be corrected. (Or if you have the DevForum post on it, please link it)

8 Likes
#34

The dev forum post complaining about the action was deleted, likely to prevent a big stink. There was another that was moderated, I believe, for having inappropriate language in comments, too.

12 Likes
#35

This is incredibly vague and frankly jarring. There’s many great snippets of code out there in the world, public or not, and tracking them down for using the F*** word as an expression in a comment is going to go down as one of the problems of this site. I am not a fan of this change and allowing for public code to be read by people we can’t trust is dystopian.

27 Likes
#36

Yes, it does. {30CHAR30CHAR30CHAR}

1 Like
#37

Do you know how long exactly ago the game was moderated?

2 Likes
#38

Couple months, I believe. November or December.

4 Likes
#39

This system would make a lot more sense if it only applied to published & public content (uncopylocked games, models, plugins, etc) instead of just blanket everything…

We’re all still waiting for the web and chat filters to be improved so that we can actually use the website and chat as intended, maybe try prioritizing those over solutions to niche non-issues which burden the user.

54 Likes
#41

this change effected my game heavily, as I use an extra chat filter to moderate my game. It replaces words like “free robux” or “f word” and replaces bypass inappropriate words. This got my game removed/moderated and my account terminated.

edit: I noticed that they silently updated this post. You’re completely wrong, roblox. I got my game moderated because of that.

this took long enough to be mesaged to us. like, you didnt tell us until now?

148 Likes
Response to code safety review discussion
#42

Does this mean the system bans users before the manual review? I thought staff would have to manually review the program to understand what the code was attempting to do. In this case, it was just creating an extra filter. (Or are extra chat filters not allowed?)

Edit: I guess my question is why were you terminated over this if you didn’t seem to break any rules? I thought someone would’ve manually reviewed it and realize “this seems fine.”

15 Likes
#43

There are a lot of questions and open-endedness that need to be addressed here, so a follow up or some kind of staff response is critical to all the concerns that are being brought up. Although it is rare, it is not unheard of for staff to overstep their bounds or leverage their authority (noting any previous incidents, such as hiding code in catalog assets) and it’s not a null possibility either.

In addition to the above, we have rightfully sensitive information and trade secrets that exist in our code that should not be accessed by anyone except developers that we explicitly give permission to but need to be there because we otherwise can’t use them. It’s one thing to enforce policies, sure, but this is highlighting a greater issue further than enforcing rules on scripts and that’s the fact that closed source code is getting read at all.

Now that there’s an actual statement about this kind of nonsense happening, it should be said: it seems extremely weird (and stupid) to spend moderation resources on things that no one except the developers see, which are better spent addressing other sectors of moderation which developers have been complaining about for several years (and still are).

109 Likes