Tavern on the Meta

@2mkgz Blacklisted user.

user259867

@SmokeDetector ignore

8:04 PM

@2mkgz Post ignored; alerts about it will no longer be posted.

-4/-4/-3 on the stats post

user259867

Interesting stats:

user259867

I'm pretty skeptical there, @rlemon. FWIW, something > 85% of spam flags are resolved as "helpful", and > 40% are handled entirely without moderator intervention... So if there's controversy here, it's affecting a fairly small portion of flags, something on the order of 16 a day for Stack Overflow. Even in cases where there's disagreement, it usually boils down to the severity of the problem rather than the outcome - only 1-2 flags a day are actually declined, with on average 14-15 disputed (but the post is still deleted). — Shog9 ♦ 7 mins ago

85% doesn't sound very reliable

the not-declined percentage isn't too bad, though

stats clean

[ SmokeDetector ] Email in answer: Minimize risk with Windows 2003 end of life by Lajohnson on security.stackexchange.com

user259867

8:15 PM

@SmokeDetector tpu SPAM

@2mkgz Blacklisted user.

[ SmokeDetector ] All-caps title: YUBICO’S TAKE ON U2F KEY WRAPPING by Lucas Kauffman on crypto.stackexchange.com

@2mkgz I'm confused how that last answer is spam...

The user asked a question in which they mentioned that they tried to contact a vendor (Stormshield). The answer was the vendor's reply.

I get that there was a lot of extra marketing crap in there, but there seemed to also be an answer.

user259867

The reply was written as an advertisement.

Seems like an edit was in order, not a "throw the baby out with the dishwater approach".

user259867

It's gone now.

8:26 PM

A baby that has been through a dishwasher probably needs to be thrown out anyways

user259867

The question said: Someone in the firm mentioned a product called StormShield... Would love to hear your suggestions... Yeah, I'm not worried about losing answers to that question.

@JanDvorak That's Stack Overflow. The numbers get a lot higher elsewhere.

@2mkgz History has shown that once a product is discussed on a question, it becomes fair game for discussion in answers, and not spam.

Drupal (for instance) has something like 99.9% helpful rate

(also 85% non-mod-handled)

(not gonna say whether that's entirely the folks in this room, but...)

rene

@TimStone quick question, slightly related to this. If I take an id from the badges tables in SEDE does their exist an url that brings me on that badge for that user? What only seems to work is this but I don't have the 1974 id in SEDE :(

8:29 PM

I'm still worried about the rejection rate on the inappropriate flag

@JanDvorak were you worried before?

There were some rejections on flagged gibberish reported here.

from before the rename

Yeah, probably gonna rename that again.

O_o?

"inappropriate" increased the use of the flag (yay) but also the decline rate (boo)

IOW, folks using it instead of closing

"disrespectful" has been suggested

user259867

8:32 PM

Offensive or meaningless

I also like "rude and abusive"

yellowantphil

"bad newz"

gibberish isn't rude

@JanDvorak is it disrespectful?

I'm wondering if most programming questions are disrespectful of the community as well

8:34 PM

@Shog9 You can play lots of games with natural language semantics, but I'd say no, not obviously so.

still a close race then

@2mkgz I've had declines by using 'offensive/abusive' as meaningless. It would be nice to know that when someone posts "hjhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh" and I flag it with that, it won't get declined.

In any case, I'm less obsessed with folks flagging gibberish as offensive since we started throwing VLQ and NAA flags into review. They'll at least get deleted without waiting for a mod now.

@Shog9 I think inappropriate is far too vague to be helpful. "Rude or abusive" would sound much better to me.

Maybe it's the description that needs changes rather than than the flag name

8:35 PM

Once upon a time, VLQ flags would sit around for hours, not really ideal when someone's dumping trash all over your site.

@AlexisKing how do you think that helps?

@Shog9 In order to understand what "inappropriate" means, you have to understand what "appropriate" means, which is not really obvious.

People know what "rude" and "abusive" mean.

The other part of this is... Can we make this consistent between post flags and comment flags?

Tim Stone

@rene I don't think so, there's a missing table of normalized badges that you'd need.

IOW, does "rude and/or abusive" work for comments too?

rene

@TimStone Hmm,ok, I'll have to craft that one myself in that case. Thanks.

8:38 PM

@Shog9 It could certainly replace the existing "rude or offensive" reason, though it could also sort of subsume "not constructive".

BTW, will we be getting spam flags for comments? If not, what's the correct choice?

@AlexisKing even better

@JanDvorak raise the rep requirement for commenting to 5000

Otherwise, "other"

@Shog9 Suddenly, the amount of NAA flags increases tenfold.

maybe

I'd actually be curious to know how much of an effect it would have

but I suspect I'm alone in that

JonasCz

Imo, the whole comment flag system should be improved (as already discussed on meta). Five choices which do the same thing ? Not very good.

8:40 PM

Should there be a penalty for repeated NAAs?

CORRECT.

@JanDvorak like... You can't answer questions anymore?

@JanDvorak [feature-request]

'Cause that one's there.

sounds good, though I meant a rep penalty

[ SmokeDetector ] Email in body: Issues Structuring Private Permissions For Cloud Applications by Robert Fine on mathematica.stackexchange.com

8:42 PM

@JanDvorak There's a sort of unspoken rule about the SE rep system that only votes, acceptance, and bounties influence rep (plus suggested edits, I guess).

and spam/innap flags

You can roll back votes, and delete posts, but everything else is pretty much set in stone.

@JanDvorak Those count as votes, really.