@bro Yes. The last I heard there wasn't a plan to change that. I'd hoped it would be changed such that posts must also not match (<code>|https?) in their body, which would reduce the fp rate by >50%.
@bro IMO, it could qualify under "abusive of the system". There's specific programmatic requirements that question bodies contain X number of characters. Just putting a bunch of ////// is intentionally trying to get around that requirement. That requirement is there for a reason. The reason is to help prevent low quality questions.
@Makyen I'm not too sure I like saying "abuse of the system" can be flagged as R/A; by that logic we can flag NAA as R/A because it's abusing the answer box
But, we're specifically directed to flag as rude/abusive posts which are cat-walks-on-keyboard, specifically because they are considered abusive of the system.
R/A should IMO only be used for extreme cases where someone seems to be deliberately trying to be rude and annoying, because of the penalties that come with it. For regular low-quality or low-effort questions, the standard mechanisms for dealing wit that work fine.
Yes, but they system requires that there be a question body, which the new user was very explicitly told, as they could not submit the question without one.
The R/A flag description is: "A reasonable person would find this content inappropriate for respectful discourse." Remember, it used to be called "offensive"
IMO using "abuse of the system" as a justification to red-flag is an abuse (no pun intended) of the wording
Yes, but the blocks are there specifically to force people to provide good quality content. That people feel free to go around those, just means that they think they are more important than what has been requested of them.
@NobodyNada I don't necessarily disagree with that, but it's what we've generally been directed to do, based on various meta posts.
I'm happy to provide the feedback that's desired by the project. I'd assumed that if our general direction from Meta is to R/A flag, then we should be providing tp feedback. Frankly, I'm not that thrilled with having the granularity of the flagging system be R/A for such posts. OTOH, it's not clear to me where the cutoff should be. The user clearly was doing something they knew to be intentionally going around an already existing system-level block (which is the general tp criteria).
@WELZ Do you want it just on the username (i.e. ^LGBTExecutioner$)? It seems that is unique enough such that it'd be reasonable to detect it anywhere (e.g. LGBT\W*Executioner).
@Secespitus-Since I have given a correct answer to your question,Please give me a bounty worth +100 reputation within 4 days from today,However this not spam at any cost,as it has correct information related to the topic,irrespective of the source — Anonymous Teacher3 mins ago
Oh right, he's the bounty guy...does he really think he's going to out-score the two legitimate answers?
What do you think about trimming a trailing "-" from feedbacks in MS? I don't think MS cares, and it prevents "fp" and "fp-" from being considered different
@NobodyNada I'm not sure I would have whitelisted that user. Circumcision is one of the John trolls favorite rants. I could certainly see him posting that with the intent of trolling answers.
@NobodyNada I'd wait for at least a revision/first answer. Given his corpus of insults, it really feels like a setup. BTW: would the question be reported again upon an answer being posted, even if the question wasn't edited?
@NobodyNada So, SD is basically stateless wrt. the posts? It only tracks ignores, whitelist, blacklist, watchlist? I would have expected that it wouldn't report the same post multiple times, without revision.
@Makyen As far as I know -- quartata can tell you more than I, but I'm pretty confident that's how it works.
Remember that Metasmoke is an "optional" extension for Smokey -- in theory, Smokey has mostly write-only access to Metasmoke. Any data that's stored in MS is unavailable to Smokey, unless it's stored on the Smokey side as well
That means Smokey only has access to blacklists and watchlists (since those are in Git), and user blacklists and whitelists, chat messages, recently reported posts, etc. (which are in pickles).
The stuff in the pickles is sort of transient, as it does not sync between instances.
OK, but I would have expected far more duplicate reports if the revision of the post wasn't checked to prevent duplicates. Basically, any question which was reported, but then was answered would generate a duplicate.
@NogShine OK. But, doing so didn't require !!/scan it again after it was already detected. Basically, !!/scan just causes a post to be put through the process which is normally done on all new posts. It's certainly been done for any post that is detected, except those that are manually reported (i.e. via !!/report).
@NogShine You can also get that information by clicking through to the MS post (using the MS link). I expect that FIRE will also show that information, at some point.
@AJ Are you really wanting a watch on that username (or some part?) in addition to that specific user being on the blacklist? That would be something to do if you are you expecting them to create multiple accounts with that username, or jumping to additional sites (if the user blacklist is per site). Are you expecting that? (I'm unsure if the user blacklist is per site, or if it's on the global account.)
@NogShine FYI you can use !!/test or one of it's variants (see the commands wiki) which will run the provided text though our spam checks and let you know if it would have been caught. It won't actually create a new report.
[ SmokeDetector | MS ] Bad keyword in body, blacklisted website in body, link at end of body, pattern-matching website in body: The creators of Burn by Whal1948 on drupal.SE
Hello, excuse my interjection. Is it possible for SmokeDetector to catch posts like these? web.archive.org/web/20180422152752/https://… these come up half a dozen times in a week. These aren't absolute spam, but I wish to know if flagging such stuff has been thought about here. Thanks.
@Zoe That pattern looks like it's already caught by Bad ns for domain in body and Bad ns for domain in answer; append -force if you really want to do that.
> Would not be caught as a post, title or username.
> Potentially bad keyword in body, potentially bad keyword in title, potentially bad keyword in username ---------- Title - Position 1-16: flexistamps.com Body - Position 1-16: flexistamps.com Username - Position 1-16: flexistamps.com
@Zoe You have 23 !!/watch pull requests already, maybe it would be about time to apply for blacklister privileges so that we don't need the PRs any more?
@Zoe I am aware of 5 admins: tripleee, angussidney, Andy, ArtOfCode, Undo. None of them are currently in this room. Other users in this room might be able to help you.
