« first day (3190 days earlier)      last day (1902 days later) » 
00:00 - 12:0012:00 - 00:00

12:17 AM
@SonictheInclusiveHedgehog Here's a site recommendation ('cuz I'm busy for a few hours), but we'll need a little more info ... What you prefer to photograph and lenses available, summer/winter for your location, is any of it at risk of expiring, and you'll need to research the look that each produces and explain which look you think you'd prefer for a particular situation. Etc.
@Rob Chat is a place for informal discussion
Throw them all away and use a 2006 watch camera.
@SonictheInclusiveHedgehog though, unless you know someone actually has experience with film or an interest, you're not going to get interesting discussion ._.
My own reaction is "Kodak does something other than <redacted> out their brand now?"
It's worth a try, plus I also posted the same on the Photography chat room
@SonictheInclusiveHedgehog so.. who here does film photography?
12:23 AM
Well, the first time I posted about it here, someone did, and we had a nice conversation
ah then its fine
Second time, same thing
12:56 AM
When I click view all review queues, the suggested edits section says 1.
But when I click on it, it says the queue is cleared.
It's not a caching problem because it happens even when I clear cache.
@forest Its a caching issue of sorts
it means there's one pending edit, but you've reviewed it already
since the count on the queue isn't generated per person iirc
Oh server-side caching.
That's silly.
"Its too expensive to calculate it per view"
so the tradeoff is between a good amount of performance and people getting confused until someone explains it to them
It's a shame SE can't find a massive community of competent DBAs who could assist with solving the performance problem for free... If only such a community existed!
@forest actually their DBAs and SRE teams are pretty competent.
and just cause there's a lot of eyes and its free dosen't mean its very good.
1:08 AM
I suppose they're competent compared to image board DBAs.
But it's always possible to find ways to improve performance.
Even if it means doing the processing client-side.
and its probably being done/tried
and SE seems to quite literally be able to tell MS "this thing is not that good" and effect change
Shame they're using MS SQL.
(in some cases)
@forest its what Joel and Jeff knew
and well, it seems to do ok
and there's a few things SE does - like clustering which might be tricky to do elsewhere.
Sure, just bug-ridden and plagued with performance issues.
But it does enough to keep the site online at least.
@forest It's a tag wiki edit. You don't have enough rep to review tag wiki edits.
1:13 AM
@SonictheInclusiveHedgehog And the site doesn't tell me this because...?
Physically-unsolvable erformance issues? :P
@forest its expensive ;p
they could do it, but its a lot faster when they don't
Q: Please don't show the red dot indicator for review items I can't do

hatThis is not a duplicate of Please don't red alert me for queues I've handled, because this reports different things from that question. That question is specifically about the case where the user has run out of daily reviews; this question reports different problems outlined in the below bullets....

Support this FR if you want it
So give the user some JSON with information about the edit.
And have the client choose to display it.
@SonictheInclusiveHedgehog Oh so that's what causes that bug!
Though I do still see the red flag issue even on a site I have 40k rep on...
It does the same global calculation for the indicator, and doesn't take into account user-specific conditions for review
1:16 AM
With so much fat, slow JS, you'd think a little extra code to parse JSON client-side would do the trick. After all, that wouldn't put any extra stress on the DB.
The actual pending edit is this one, which is a test edit; I don't think there's a need to keep it pending anymore since the corresponding report has already been marked as under review by an employee.
@SonictheInclusiveHedgehog OK, that explains it, NM.
1:33 AM
You know, a lot of spam I never end up flagging because I don't want to register to another site on SE. It would be a lot easier if that weren't the case.
For sites with a totally different domain, I understand why there'd be limitations.
But when the site is just another subdomain under stackexchange.com?
Some people just don't want accounts on other sites.
No need to create an entire new account.
Just give very basic privileges like the ability to flag as spam.
I.e. things which don't require you register a user and profile.
Flagging requires 15 rep
Well yeah and flagging also requires an account. My point is that it shouldn't. :P
If I already have an account on the network, I shouldn't have to sign up to a new one just to flag a single spam post I see.
Only a small proportion of users actively need to go to other sites just to flag spam
1:40 AM
No one needs to. My point is that I have skipped on flagging potentially dozens of spam content simply because it's on a site I haven't yet bound my account to.
2:09 AM
@forest yanno what would be a great way to break the flagging system?
flooding it with flags
the need to register (and getting banned for bad flags) make that harder
Naturally I don't think unregistered guests should be able to flag.
But once you already have a network account, it shouldn't require anything more.
and don't forget, if you have an account on one site with over 200 rep you can already flag
and a 'new account' is 2 clicks
Yeah but I don't like having to keep track of each and every new account.
It screws up the metrics page.
I mean I guess I could create a new account then delete it.
hm,would be messy
or have a sock for that.
getting 200 rep on some site is trivial
Why would it be messy? Does account deletion not nullify the account totally?
2:13 AM
ya but you're generating an account per flag
which eventually is going to get someone annoyed
Why would anyone care?
you'd be surprised
We're thinking of how to suspend that user if the flag is abused
Well naturally abusing the flags is not what I had in mind. :P
People do all sorts of stuff that annoys people that seems trivial or mechanically fine
2:14 AM
Naturally, I always think of the worst case when implementing a feature...
always a good idea
@JourneymanGeek If I create an account (well, "join a community") to raise a flag, what's the downside to later deleting that account? Isn't it usually automated?
@forest doing so for each time you want to do it...
seems excessive
To whom?
Is a mod forced to watch a video of Rick Astley for each time an account was deleted?
Well, personally I feel folks ought to be engaged in a site if they want to flag, smokey aside (and I've probably thrown a fair bit of hopefully constructive critique their way)
2:18 AM
I was thinking more about spam where there's no use in providing critique.
and "why would anyone care" is a poor justification for anything
@forest well, we have a pretty robust way of handling it...
That's true.
I've always been impressed at how quickly SE deals with spam.
I even told a spammer once not to bother with non-directed spam on SE because they'd be wasting their time badly. :P
2:58 AM
@JourneymanGeek Excessive how? I thought deletion was simply that. Deletion. And association bonus definitely makes your life easier in this case.
@Mr.YouKnowWhoIAm actually that's one possible thing
Whether you want to blame the system or not, it is hard to audit deleted users. There's even a troll abusing the account deletion feature by posting a troll answer then delete the account immediately.
that too
That sounds like a SE problem.
@forest When you mean quickly, how quickly? 2 min., 5 min., 10 min. How many minutes does it take for a spam post to go bye-bye?
3:01 AM
(sorry, distracted, Trying to troubleshoot a printer that's shooting blanks)
@Mr.YouKnowWhoIAm 30 seconds
@Mr.YouKnowWhoIAm For questions, it often takes mere seconds.
not even kidding
For answers, it's rare to last more than 10 minutes.
even without smokey - on fairly active sites, I've seen sub 20 seconds to first flag
Though sometimes people remove the spam links via an edit and leave the answers, then they can last for years and spam flags get declined...
3:03 AM
@forest @JourneymanGeek That’s speedy. Give me one reason why.
@Mr.YouKnowWhoIAm cause people care
Because hundreds of people instantly see an obviously spammy title.
also SE is designed in some ways to reward that kind of behavior
Plus auto-flags make it less difficult to get rid of the posts.
In combination with the rather hardworking folks who do the... charcoal thing.
@forest Since when does spam have to include spam titles? There is something called subtlety, you know.
3:06 AM
Most spammers are not very subtle.
After all, they want you to see their spam.
@forest The best way, in my opinion, is to trick you into looking at it.
Sure but the best spammers also do targeted campaigns.
They avoid SE specifically because it's not an effective medium.
(Disclaimer: I was briefly a spammer for what's termed "blackhat SEO" before I quit due to not wanting to support the capitalist machine, so I've talked with a lot of those folk)
@forest You know, if there was a lot of spam (making Smokey go cuckoo), it would take a long time to get it all out.
They're pretty much exclusively forum, blog comment, and email spammers.
I'm not surprised by either.
@forest we get the worst spammers.
3:09 AM
@JourneymanGeek You get dumb spammers.
and not ones for german meat products, which would be the wurst spammers.
@forest we do
and I'm convinced they literally use 3-4 standard toolkits
They do fuzzy matching for posting so it works on many forums.
Pretty much always clones of XRumer.
@Mr.YouKnowWhoIAm there's already SpamRam built-in on SE system that already prevents most spam being posted though
(there's enough commonality - and the fact that they use white person name generators written by someone who clearly has no names that a actual white person would use)
But, yeah, we're currently helpless against mass username spam
3:11 AM
@JourneymanGeek They use prebuilt collections of names and phrases.
Please elaborate on “blackhat SEO” (reminds me of blackhat hacker), capitalist machine, German meat products, standard toolkits, XRumer, SpamRam, mass username spam, etc.
voted to close as Too Broad
Blackhat SEO is using stuff like spam and meta tags on spun (auto-generated) blogs to increase the ranking of certain search terms for marketing purpose.
Capitalist machine is just the term I used because I hate capitalism.
XRumer is a popular spam tool (botmasterlabs.net/xrumer).
Please elaborate on the term “capitalism”.
@Mr.YouKnowWhoIAm Basically Might Makes Right, but applied to industry.
3:19 AM
Well, SmokeDetector makes things a little faster in terms of spam.
Fun fact: XRumer popularized spam seeds (which SE knows all too well about).
(Not something I knew before, I had to look that up lol)
Even ReCaptcha is pretty useless against these bots.
People can buy dirt cheap captcha solving services.
(Which are actually pretty useful even if you're not a spammer)
The captcha gets sent to a remote server and the server uses OCR and NNs to try to solve it on its own, and if it's not confident then it'll send it to humans to solve.
Honestly the best solution is a vigilant community.
Which SE actually does really well.
@forest Hmm... including the captcha that has squares that make up one whole picture and tell you to find the traffic lights, fire hydrants, etc.?
Though those are actually not that hard for computers to solve.
You know the real reason Google uses those?
They're training their own machine learning algorithm that they sell to the US military.
Yes, that's actually what it is. The Intercept wrote an article about it.
I was going to say, “for collecting Captcha data?”.
So yeah, it's evil of them, and not even all that effective.
The real reason it has any effectiveness is that it's slow.
Even a legitimate human often takes a lot of time to solve it.
3:26 AM
if you wanted to be super sneaky
something solving a captcha too fast would be a givaway
There are a lot of heuristics it uses to detect dumb bots.
But smart spammers will either use a solver service, or proxies.
Vip72 is a common proxy service, or they just use small router botnets.
I prefer router botnets just because the rDNS is less suspicious.
Whereas I've seen a few corporate networks for Vip72 that say "proxy" in them.
@forest The articles you have given me were dated approx. 1 yr. ago!
@Mr.YouKnowWhoIAm So?
@forest Think of how much reCaptcha data they have collected now!
Yeah, that's why they're moving towards grainy pictures.
Which look suspiciously like it's trying to emulate high-altitude aerial footage.
3:29 AM
@forest Grainy? As in blurry?
As in grainy, like with noise. Fuzzy, not blurry.
I'm skeptical, but I'm lazy to ask about it on Skeptics
@NoDistractionWizard About the aerial footage, or the whole thing?
Because Google has admitted to this.
There was even an open letter from dozens of employees fighting against it.
The whole "looks like it's trying to emulate aerial footage" is just a guess of mine though, but it certainly seems like that.
@forest Ah, even more effective power!!!
@Mr.YouKnowWhoIAm More like low-light conditions.
@NoDistractionWizard There's a reason I hate humanity. Most people probably consider Ed Gein or Roy Norris to be evil serial killers, but I consider the same for anyone who willingly uses Google reCaptcha. You support mass-murder and genocide.
3:32 AM
I meant, the harder identification + the abundant reCaptcha data => even better AI!
@forest Remember, it isn’t just for military. It may as well as be used in everyday life.
You think Google's ML research is public? :P
No, but it may help after some time.
I don't know about you, but I personally feel that genocide is not acceptable, even if it gives us a slightly more accurate Google Street View.
Haha. Funny. I am not amused at all.
3:37 AM
Wasn't meant to be a joke.
You forgot the last sentence.
No I didn't.
Remember, self-driving cars.
The sad thing is that the original reCaptcha was designed to help digitize old public domain books for the public use in libraries. Now it's used for murder.
3:40 AM
Please elaborate.
@Mr.YouKnowWhoIAm google used to have a project to scan ... everything they could
@Mr.YouKnowWhoIAm You remember that captchas with one word that looked warped, and one that looked like it came from a scanned book?
the idea was to make real books searchable.
You had to enter two words. The warped one was generated by a computer, the scanned one came from a book (and actually was a no-op, so you didn't have to correctly solve it, just put in any text, but most people copied it faithfully).
3:41 AM
and recaptcha literally was initially human managed ocr
@forest i figured they took a bunch of values then used... statistics.
For the scanned one? It just collected what you wrote.
4chan used to have a campaign (which failed) to enter in millions of captchas, replacing the word "the" with the N word in the hopes that public domain books' digitized version would occasionally replace "the" with racial slurs. For the lulz.
It didn't work though due to the sheer amount of correct answers.
@forest 4chan isn't really that many people though.
You'd be surprised.
They've gotten huge things to change before, including the TIME person of the year.
And their counter-raids on Tumblr were legendary.
Those threads were hilarious.
3:44 AM
4chan sounds like something I won’t like.
@Mr.YouKnowWhoIAm 4chan has a lot of different boards, all are different.
Some boards like /pol/ are obscenely racist. Others like /a/ are quite pleasant.
^ This is one of the most amazing things they pulled off, recently too.
> This is when their man on the ground came into play. The channers were able to enlist a local troll to drive around the area and repeatedly honk his horn. I would like you, my fair reader, to imagine a man driving around a small Tennessee town in, what I assume to be, a 1994 Toyota Tercel just blaring the horn repeatedly while rocking a, again I assume, soundtrack from a Metal Gear Solid game, all in an attempt to fuck with Shia Lebeouf—this is a thing that actually happened this week.
And more. It was remarkable. I watched some of those live streams. :D
They even used star charts and airplane trails to geolocate a single flag.
@Mr.YouKnowWhoIAm But in general, it's /pol/ and /b/ that people hate.
I stopped reading as soon as I saw despised words.
@Mr.YouKnowWhoIAm Shame. It's a remarkable story.
> The trolls using only the live stream of the flag, started, I shit you not, studying the flight patterns and contrails of the airplanes passing overhead. They mapped out what they saw and took their findings to flight radars to try and pinpoint a general area. Using the knowledge gleaned from the flight patterns they found that the location was near Greeneville, Tennessee.
And then used echolocation with a car horn to finally locate it on the live stream.
All to replace a flag in a random field with a hat. :D
erm. Man, don't these people have (other) hobbies?
I am terribly confused. Anyways...
3:49 AM
@JourneymanGeek Who /pol/? Yeah they also screw with Ben Garrison.
anyone who would work out where a flag is based on airplane flight patterns, or car ecolocation
But it's pretty much limited to screwing with people to see their reaction, and making racist jokes. It's genuinely hard to tell the difference between actual national socialists and trolls on that board. Just remember... not all 4chan is like that.
@JourneymanGeek Oh it was a group of people, not just one.
4chan can be amazingly coordinated when they try.
I do mean collectively
Other times... they can't even count to 10 without C-C-C-COMBO BREAKER.
(Note: I am an /a/non, not a /pol/ock)
Which makes me sad when everyone thinks 4chan = /b/ and /pol/.
Those are literally the two most despised boards. Even other 4channers hate them.
archive.4plebs.org/pol/thread/115783569/#q115789191 lol this is when they found the airplane.
It's been a long time since I visit 4ch...
3:54 AM
Like freaking private eyes.
I gotta respect them.
Do you... like Shia Lebeouf?
@forest it dosen't really matter
I mean, of they hadn't done it, I don't think anyone would have remembered his 'art project'
or better yet, thought of him as a nutter
That's probably true.
2 hours later…
5:45 AM
@forest They must view the sweater instead.
@JourneymanGeek, that was funny, German spammer.
1 hour later…
7:05 AM
So this morning...
apparently a lightbulb in our cooker hood exploded
tripping the breaker, so I got woken up when I was planning on sleeping in
ended up removing said grease covered hood and popping in the designated bulk waste disposal
also trying to fix a printer
today was productive?
@JourneymanGeek Depends on whether sleeping in was a task to be done too ;)
though with the printer
apparently the step I'm at is purging the ink tubes, then waiting 12 hours
@JourneymanGeek waiting 12 hours means time for sleep, right? :P
@Tinkeringbell well I got to that part at 11am...
and I don't sleep in the day time unless I worked nights
Ah. Well, then it's time to get rid of a principle :P
3 hours later…
10:22 AM
1 hour later…
11:34 AM
Point of order - on a personal level. The above comment kinda ends up dismissing Tink's answer, even if it wasn't intended. Its worth considering folks do spend their time and energy writing answers and going "But I'd rather have one off an employee" seems dismissive. — Journeyman Geek ♦ 7 mins ago
@JourneymanGeek Deleted and reposted without mentioning employees
Oi @ShadowWizard, taking it to chat not to clog the comments there. You meant the dummy edit from two days ago? I didn't suggest one today AFAIK
Hi all, by the way :)
@SonictheInclusiveHedgehog its worth keeping in mind in future too.
@Jenayah #nocontext ;p
Ahah yeah they'll know
@Jenayah oh whoops... maybe! Didn't see the date, but not used for edits to stay for so long. lol
@Jenayah Comments where?
(so others know)
@Jenayah I don't
11:36 AM
@SonictheInclusiveHedgehog no comments now, from me at least. ;)
Here meta.stackexchange.com/q/327373/398063 but some got deleted
@ShadowWizard no worries. Had me worried that something iffy was going on, that's all :)
@Jenayah Thanks. FYI, some users were bothered by the review indicator caused by your as-yet-pending edit.
10 hours ago, by Sonic the Inclusive Hedgehog
The actual pending edit is this one, which is a test edit; I don't think there's a need to keep it pending anymore since the corresponding report has already been marked as under review by an employee.
@Jenayah oh... and here @rene meant it's working fine in Chrome. but in your question is appeared as "not working in Chrome"
(I think? If I misunderstood rollback, but better clarify)
@SonictheInclusiveHedgehog rejected
@SonictheInclusiveHedgehog oh shit I saw that as well, a pending edit which wouldn't load, and was convinced it was a tag wiki edit since I still wasn't 5k. Auto-trolled :D
@JourneymanGeek To be clear here, is the complaint about the fact that I expect answers from employees, or that I expressed that fact in a dismissive way?
11:40 AM
@ShadowWizard ah, I guess I thought the blank excerpt thing was a bug
@SonictheInclusiveHedgehog To be clear? I'm not going to say yes. I'm going to say both are true
@SonictheInclusiveHedgehog you can't demand answers from employees only, for that we have the contact form. By posting on MSE, you also "agree" to get answers from anyone, even if not employee of SE.
Well actually, isn't it a bug? It shouldn't take my dummy edit into account, but still have the current excerpt?
Do non-employees have access to the code?
@Jenayah hmm.... yeah. Good point.
making another coffee
11:42 AM
@SonictheInclusiveHedgehog the thing is though, you're expecting them to specifically look at your posts
dismissing the collective knowledge of the network.
@SonictheInclusiveHedgehog not really. So?
And there's stuff some old timers here know employees might not
or as mods, we might know something about some mechanics and not others
@ShadowWizard ahahah
or even random stuff we happen to pick up off twitter
That's true. Employees can't know everything, but if they don't know something, they can access the code and check.
11:44 AM
Assuming they have the time, inclination, and access
@SonictheInclusiveHedgehog they can still do it even if someone else already posted.
@Jenayah actually, the blank box bug was reported and fixed long ago:
Q: Unable to improve tag wiki excerpts (I get a blank box)

GillesI was reviewing suggested edits on Stack Overflow and saw two tag wiki excerpts that I wanted to improve: source-separation, convex-hull. When I clicked “Improve”, I reached the usual improve UI. However the box that should have contained the text of the suggestion was blank. At the time I click...

So it somehow floated back. :/
Ghost bugs... Eventually them come back to haunt you
@JourneymanGeek Nowhere did I say that there's anything wrong with people making their best efforts to figure out what is happening in specific cases.
@ShadowWizard ... did I miss something?
11:50 AM
@SonictheInclusiveHedgehog that's implied though
as is the possibility that if someone found the problem/bug interesting they would dig.
Having an internal preference for staff answers, or expressing that it would be nice if staff answered, is not implying that community answers are bad.
@AdamLear marked 18 (!) questions on MSE as in the last two days. So it means SE now appointed a full time developer to work on bugs posted on MSE? Pretty sure that was never the case.
@ShadowWizard and/or their new PM is grabbing that
hmm.... hopefully.
@SonictheInclusiveHedgehog do consider the feelings of the non staff folks who answered
11:54 AM
Is he pingable here? @Ces... nope. :/
oh wait, he's a CM. Who is the new PM?
@ShadowWizard her
I'm terrible with names
@JourneymanGeek "Cesar" is a male name, at least where I'm from.
@SonictheInclusiveHedgehog but he's not who he meant
11:55 AM
@SonictheInclusiveHedgehog PM not CM
42 secs ago, by Shadow Wizard
oh wait, he's a CM. Who is the new PM?
Ah, I missed that.
hehe what a mess. Let me search then....
> Megan Risdal 👾
11:57 AM
@JourneymanGeek Yeah, perhaps putting it in a comment wasn't the best idea, and now that I think about it, it does briefly give off the appearance of being dismissive.
Good morning! So excited to share I've joined @StackOverflow as a Product Manager! Couldn't be more thrilled! 😊 #FDOW
@SonictheInclusiveHedgehog it does
@ShadowWizard the odd thing is I remember handles fine
Normal People names on the other hand, kick my tail
00:00 - 12:0012:00 - 00:00

« first day (3190 days earlier)      last day (1902 days later) »