« first day (1817 days earlier)      last day (1712 days later) » 

00:07
I kinda like their use of underscores instead of parentheses with multi-word alternatives. I'd never seen this before. Definitely useful, but somewhat unsightly.
00:26
@userr2684291 Eh? I think you yourself used underscores?
I think it's favored for prolific SE writers. Harder to miss in the text area
@userr2684291 @M.A.R. I have to star both. Both are great.
@userr2684291 yeah when you were as hot headed as our young fellow here
Strokes white old gentleman beard
@userr2684291 this is definitely how I word it on sites that are stricter about their rules
On an ELL that flagged my welcome comments as unfriendly and facetious, I'm not sure it would fly for long.
Most likely, the commenter will become the unhappy lifeless jerk everyone has learned to ignore
AIQ
AIQ
@M.A.R. Really? What was it?
That was it. I dunno how to explain further. It's ancient history though
Nothing too dramatic, but it evidently rubbed me the wrong way
01:21
Hi
Can someone help me?
"If player does not have the first X, > he < will not be able to select T"
is "he" correct?
I should use "it"?
translate tell me use "he"
:((
 
1 hour later…
02:34
@SinNombreSinApellido Correctness depends on context: in this case we can focus on the audience (people reading your text). For some, there's nothing wrong with the genderless he, but others perceive it as limiting the mentioned players to men (rather than including women), especially given that more inclusive alternatives exists. E.g., you could use they, or he or she, or he/she. It isn't used to refer to people, unless they're babies (some people find this objectionable).
02:47
Note that this is just how most people write and say it, so that's popular usage. I wouldn't go beyond that myself.
@SinNombreSinApellido It is a sexist translator, either a human being or a piece of software designed by sexist humans
@M.A.R. Nope.
 
1 hour later…
03:58
Word of the day: arthritides
What is the medical term for the drug side effect in which your breast increases in size
04:25
> Pleural effusion was more common in patients with CML in accelerated phase and blast crisis as compared to patients with chronic phase CML (over the observation period of 1 year).
THE or A?
> These potholes are a historical monument. Under protection by Yaroslavl's City Council.
 
2 hours later…
06:43
[ SmokeDetector | MS ] Offensive body detected, offensive title detected (95): (potentially offensive title -- see MS for details) by Dmitry on ell.SE
 
5 hours later…
11:16
@EddieKal I wouldn’t go that far. It wasn’t that long ago that using “he” as a “neutral” pronoun was the rule. It’s good that it changed but I have a little more compassion for the people that learned that rule when it wasn’t recognized as sexist.
I don’t think it’s constructive to assume we know what is in stranger’s hearts.
11:28
@CowperKettle I might write something like "observed over a year" but I don't know that it fits in that particular sentence. I would use "an observation period", not "the" in that one.
@CowperKettle You could also write "over a one year observation period"
 
2 hours later…
13:01
In all fairness, you're in a language learners site. Learners are never ever taught singular they or the sensitive ism issues.
In fact, our "TOEFL preparation" textbook, which the teacher chooses exercises from, explicitly has things like "Someone is doing their job" which we're supposed to correct to "Someone is doing his job".
@M.A.R. Yeah, I think that it's important to point out to learners that can be perceived as sexist by some people, but we shouldn't call people names because they were following a rule they were taught when may not be aware that it's changing.
@ColleenV It's not even "may not". I've yet to see a learner's course mention this
99 percent of learner don't know about this. 100 percent of those who don't know default to "he", even when they're female
80 percent of stats are made up on spot, but mine isn't
@M.A.R. It's better to add a decimal point when making up stats. 78.4% of people find them more convincing.
13:20
@ColleenV and sometimes you can be wrong theguardian.com/world/2020/jun/05/… This was reported on Friday and elsewhere: Martin Landray of Oxford, a leader of the trial, told Reuters “This is not a treatment for COVID-19. It doesn’t work”
[…] And no, I don’t think that there is a sudden switchover to efficacy when zinc is added, but you know what? We’ll be getting trial data on that idea, too. (Derek Lowe)
@Mari-LouA Well, it's not a cure
Goddammit will people stop calling it that. Ugh.
@Mari-LouA And I'm not wrong when I say you should be skeptical of people making claims pro or con about it. That article right on the heels of the other studies being debunked just supports my point
We don't know with any certainty how effective it is or isn't for COVID
Hmm, yeah, it didn't seem as certain by the time I finished it
It's pretty easy for a sample of 1500 people not to be random. Maybe there's a statistical error in the original studies that show it's effective, maybe one in the latter when they say it doesn't
Well they don't have a link to the paper, so there's no way of judging the conclusions they drew from it.
> “One of the key lessons we should learn historically is that making treatment decisions based on observational data is not the way forward,”
13:29
When the pandemic began, I read some studies on it to get some preliminary knowledge on what it is and isn't. I haven't read any studies since then TBH.
@ColleenV And that's exactly what they did
AIQ
AIQ
-> "That article right on the heels of the other studies being debunked"
@AIQ Sounds painful
AIQ
AIQ
I think I get what this means but then I don't ... is something missing?
Why can't everyone just get a chair
@AIQ "that article being published right on the heels of the other studies being debunked..."
AIQ
AIQ
13:32
well ColleenV is talking to you and Mari - you both understand that, but I am not at that level but I liked that part - but I don't quite fully understand the "right on the heels"
Meh, the hype is political, the debunks now seem a bit political too
Idiom of the day : on the heels of Following close behind or soon after someone or something.
Well that link is borked now
AIQ
AIQ
oh thank you
now that makes sense
Adding "right" just means "really soon after, almost immediately after"
Think about someone following you so closely they step on the back of your shoe
AIQ
AIQ
13:33
yeah, that was nice to read, and now I steal it
Word of the day: Consanguinity
2
@M.A.R. When things are this political, you can't be certain of anything
A complicated word to express a simple idea. Use with caution, and to confuse people
AIQ
AIQ
@ColleenV [that article] [being published right on the heels of] [the other studies being debunked] [just supports my point] Now I get it - I had to break it down
Well done.
There's also "hard on the heels of" which is similar to "right on the heels of"
Hard has a lot more meanings than I normally think about ... you can give someone "a hard look"
be a "hardened criminal"
cause "hard feelings"
be a "hard worker"
AIQ
AIQ
13:42
a hard look
be hard of hearing
A detergent can be "hard"
ooo! We can do "hard science"
AIQ
AIQ
A hard drink
hard = alcohol
And those are just some of the adjectives!
AIQ
AIQ
Hard water contains a high level of minerals that prevent soap from cleaning.
> The economic downturn has hit the poor especially hard.
AIQ
AIQ
13:45
my country has made 3 zones - red, yellow, and green for the COVID19
The adverbs seem to be similar to the adjectives
> 5 : close in time or space stands hard by the river
@ColleenV And by all accounts, it's neither a reliable treatment for COVID 19.
That's what I was looking for to explain "hard on the heels of"
AIQ
AIQ
jesus
@Mari-LouA You can believe that if you want. There's a chance you might be right
Your certainty is unwarranted though.
 
2 hours later…
15:32
@ColleenV I know I am happily perched on one end of the spectrum with my heels dug in, but I agree with you that making assumptions about other people is not the best approach. I was half-joking, making a half-serious statement with the other half.
@EddieKal I try to focus on the action being whatever-ist instead of the person
It rarely helps a person do better actions if you talk about them as if their bad actions are a result of who they are rather than just a mistake out of ignorance or poor judgement or what-have-you
And, I personally believe it damages my thinking when I label people so I try to avoid it
What is really, really annoying? Clue: The difference between earlier and before. Why is it that no one around here, especially among the native speakers, doesn't come out and tell certain posters that earlier and before do not mean the same thing??
I think I saw someone get it right earlier, before you mentioned it ;)
i share your frustration though when people don't see differences between two related-but-not-the-same words
15:57
@ColleenV Well, fumble fingers sure had a field day before the question was closed....
16:47
[ SmokeDetector | MS ] Offensive answer detected (78): What does "contract a trace" mean? by Eugene on ell.SE
@Lambie I used to frequent ELL's home tab earlier before, but now I don't
And I'm not a native speaker clear conscience
17:45
@M.A.R. Ah, right. By the way, can you explain to me why that smokedetector bot places what it detects as offensive here?? Also, I read the question and saw nothing offensive.
@Lambie Smokey's (as we call him) main room is Charcoal HQ, where it posts what it suspects to be rude/offensive or spam for people to review the post for flagging. A couple of years ago we requested it post potential rude/offensive stuff from ELL here as well, so if there are active mods or users in chat they be alerted quickly about it and do whatever needs to be done.
If you're referring to the c-word question, it's obvious why it suspects it might be offensive, but it's definitely not a final verdict; it's an algorithm that mass-searches posts with recent activity.
IIRC it was me who requested it after a small poll here to make sure people are okay with it. If it becomes bothersome turning it off would be a simple matter of pinging the right admin to toggle a setting
@Lambie Probably the combination of "perverted" and "fascist" or something
Oh that one, heh
If you click on "MS", you'd find out the reason the post was detected
> Offensive answer detected - Position 126-133: fascism
It's not a particularly offensive word, but I think the trolls that pop up in religion sites or politics.SE might have something to do with it
18:01
I fear we are moving more and more into the realm of what is PC when bots and their keepers are clueless. This is the kind of thing that got you-know-who in trouble. In literary terms, it's called misprision. :)
Eh, I don't think we're in that danger. It's just a bot, it doesn't do anything
People don't flag posts because they contain the word "fascism", and indeed that's not a logic to conjure such an algorithm
Say, some troll went to Judaism.SE and said "Fascism is great." yadda yadda
In fact, no need to speculate. I'm gonna search the caught posts and see why such a detection reason was added. Sometimes it's just a pattern that happened thrice back in 2018 and is still in the codebase
That seems... less than useful
Mhm, it happens to be a common enough keyword in troll answers to legitimate questions, but it throws some false positives too
Not a big deal for a small site like ELL I think
It doesn't DO anything without humans actually flagging, right?
It's never a big deal. Detecting bad apples at the cost of some false positives is much better than not detecting them at all. Without Charcoal, spam would sit around for hours, if not days, at quieter sites
@ColleenV Smokey autoflags posts that pass a very high "reason weight"
99.95% accuracy at least, I think
18:12
Right, but that pattern doesn't
Even then it's not six flags.
Or does it?
@ColleenV Sure it counts, but its reason weight is 78
78 percent of answers with 'offensive keywords' have been true positives
To compare, spam posts can have a reason weight of 1200!
Smokey "autoflags" them four or five times, I forgot, and the sixth flag would be a human review
Autoflag is really users that have registered for this function spending their account's flags
We have "preferences", like "only use my flags for reason weight above, say, 250, not including such and such reason"
That's at least three good detection reasons
Charcoal is not often accused of overt PCism, but there's always a group of users that always look at organized moderation in a negative light, like a clique that deletes 'whatever they like'; as such, they have always been under extreme scrutiny
I guess at least until people found someone better to be pissed off at
 
2 hours later…
20:30
word of the day: precariously
3
 
2 hours later…
22:28
@M.A.R. Accuracy is a bad metric here, if you're talking about the ML formula.
I believe the AUC and F-measure performance metrics are what's used with classifiers for imbalanced datasets.
In software defect prediction there's been some progress recently in this regard with some hybrid models involving neural nets. On average, you can get up to 85% of whatever these more accurate metrics represent.
I, um, informally collaborated on some research regarding this.
I'm not really interested in the matter, though, lol. I was kinda forced to, and just did it, I guess.
22:52
Hm, although I am assuming here that there isn't that much spam, and that we don't have the data... I might be wrong, lol. I guess I see everything through SDP-colored glasses.
Email spam and regular internet spam (forums and other "boards") could be used here... yeah. Whoops.

« first day (1817 days earlier)      last day (1712 days later) »