5

I'm sometimes accused of being a bit heavy-handed with my closevotes, but at least if I get it wrong, there have to be four other users voting the same way before anything actually happens. Whereas pseudo-user "Community" is right up there with the moderators in terms of being able to act unilaterally.

But in relation to What is the most common opposite word for “postpone”?, I'd just posted my own "duplicate" closevote to the question when "Community" unilaterally deleted this ANSWER...

The most commonly used antonym for postpone is advance. However in the context of that sentence, the sentence would have to be restructured to use the word advance. You can check for great synonyms and antonyms on the vocabulary building app [link removed]

...citing as justification...

This answer was marked as spam or rude or abusive and is therefore not shown - you can see the revision history for details.

Can anyone explain why Community did this? Is it something automatically triggered by the link? Is that on some kind of "blacklist"?

8
  • 1
    Well, actually, I am deleting my answer because I am a bit confused too. Usually when SmokeDetector picks up a potentially problematic post, it automatically leaves a flag on the post. I don't see that flag. The only flag I see is from the moderator who deleted the post. I think they probably saw the notification somewhere, most likely in a chatroom. SmokeDetertor pushes notifications to a lot of places. I am only aware of its notifications in the ELL main chatroom and here is the notification regarding that answer. – Eddie Kal Mar 21 at 16:36
  • 2
    So my answer is not an accurate explanation of what indeed went down. I think we should wait for that moderator or someone else with better knowledge of the workings of SmokeDetector. – Eddie Kal Mar 21 at 16:40
  • Yes, I realise that some people might want to see this explored in more depth (on another day, I might too! :) Broadly speaking I'm happy with the basic principle that SO should be relatively "unforgiving" in regard to links whose primary purpose seems to be "promotion" (but obviously we need to be able to link to dictionaries all the time, and not infrequently, querents and possibly others might need to link to "teaching aids" containing questions, examples and other relevant text). But if I see such a link in future, in an answer which has merit in other respects, I'll edit out the link. – FumbleFingers Mar 21 at 16:57
  • (I'm not after criticising the other mod, but if it wasn't too onerous to implement, I think I'd have much preferred that only the unwanted link and immediately-preceding "recommendatory" text should be deleted. But maybe it's unreasonable to expect moderators to give that much attention to posts that are usually completely "unwanted".) – FumbleFingers Mar 21 at 17:02
  • Handling of the answer [MS] was probably informed by the fact that the user has a history of spamming for that domain on SE and their posts promoting that app/domain are their only contribution to Stack Exchange. The user is also promoting that app/domain on multiple other sites on the internet (see search link in first comment; not repeated here in order to not add an additional reference to the domain). – Makyen Mar 21 at 18:49
  • Handling of posts with that domain is on a case by case basis. There are three posts on Stack Overflow which use two different subdomains which were considered to not be spam. – Makyen Mar 21 at 18:51
  • 1
    As to editing out the spam link: Editing spam is harmful for multiple reasons. If you feel what the spammer has said is beneficial as an answer (without the spam), then it's better to post your own answer (in this case, on your proposed duplicate-target), but double-check that the spammer didn't plagiarize the non-spam content (which is a fairly common tactic for spammers). Admittedly, it is sometimes hard to determine if the user is actually spamming or if they are a well-intentioned user who is linking too much. – Makyen Mar 21 at 18:55
  • 3
    @EddieKal SmokeDetector and metasmoke only automatically raise flags on posts which are very likely to be spam. From a moderator's POV, autoflagging can be recognized by having a spam flag from the SmokeDetector account, as such are only, and always, raised when autoflagging. Users can set the conditions which they desire for when their flags are used for auto-flagging, but are not permitted to set any condition which is less than 99.75% accurate, based on the historical accuracy of how that condition would have performed on all detected posts and the feedback from humans which was received. – Makyen Mar 21 at 19:01
4

I flagged that answer as spam, and since ♦ moderator votes are binding, one of those flags is enough for the Community user to delete and lock the post, as per the FAQ.

Is it something automatically triggered by the link to beebl.io?

You're close; that specific link triggered SmokeDetector because the domain has been seen before in spammy posts. I saw a message about it in one of the chatrooms where those posts are reported.

I could have chosen to edit out the link instead; I sometimes opt to do so if the content is particularly valuable. In this case, I think our regular users are able to post a better answer, e.g. with references and/or a concrete example.

1
  • 3
    I'm more than satisfied with this answer, thank you! The "non-spammy" part of the deleted answer wasn't that valuable, and seeing you explicitly say that if it had been, you might have edited to preserve the parts that did have merit is quite enough to convince me the system works well. (So much for me thinking I might have found a bug! :) – FumbleFingers Mar 21 at 17:39

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .