Two HN Announcements
blog.ycombinator.com
blog.ycombinator.com
That might scare some people away from vouching. Could you clarify whether it'll be more like, "If you wrongly vouch for even one single thing, we'll silently and permanently remove your vouching ability forever with no possible recourse" or more like, "If you show a repeated pattern of bad vouching, we'll reach out to you and explain what you're doing wrong, and only if it continues, take away your vouching privileges as a last resort, perhaps only temporarily" (or somewhere in between those extremes)?
P.S. I couldn't be happier to hear about Dan's promotion. He has an expert touch for community management, and (I learned after an opportunity to join him for beers one night) some deep wisdom on the subject, too.
I wouldn't have even included the bit about taking away vouching rights except I know that the question "What if people just vouch for all the bad comments" was going to come up otherwise.
(Also, I don't think I've been promoted? But thanks—that's particularly meaningful coming from a seasoned veteran of the early Reddit...)
Editorial independence is a big vote of confidence; they don't toss that set of keys to just any old schmuck.
But I agree, this is a huge vote of confidence
And will YC companies still get what appears to be preferential placement ("appears to be") for stories? [1] Some "more equal" than others?
If this is true independence then those benefits would be extended and/or eliminated.
[1] Or do stories about YC companies simple get more upvotes because of the balance in the community here toward people working at YC companies?
If you want to make claims about preferential placement (and yes I realize you're just saying "appears", but still, people are quick to believe these things), I'd appreciate concrete links so we can look into them. We try hard to be even-handed and when there are marginal calls, err on the side of not playing favorites.
YC founders and current/former employees are a large and valued part of this community, stories about YC startups often are interesting (by HN's definition), and lots of HN users like to follow them in particular, so of course you'll still see plenty of YC-related stories on the front page. It would be weird not to. Of course those stories are subject to upvoting and flagging as much as the others are.
Best wishes.
That's a little uncharitable, considering this feature is acknowledging that the moderators can make mistakes.
It seems to me the whole point of the vouch features is to have an easier recourse for other commenters to respond to people being shadowbanned (or otherwise dead for mistaken reasons, like tripping a filter or unfairly flagged by other users), rather than actually having to find the mod email and send an out of band message.
My flag link disappeared one day without notice or explanation, and stayed disappeared for a year or so. I continue to fear using the flag link, even when I think something should be flagged. There's no chance I'm going to vouch for something other people have flagged: I'd never have enough certainty that I'm more right than they are in your eyes. I value my ability to participate in this community too much to help moderate it under threat of punishment for doing so poorly.
Re: the replies below, the worst that can happen is not losing the vouch button, it's being silently shadowbanned for something else, when that wouldn't have happened if you hadn't put yourself on the admin's vouch review list for extra scrutiny. I already fear that happening any time I participate in one of those "what are you working on / what are your side projects" threads and include a link to my site.
This is confusing to me. Unless there's something nobody told me (always possible!) the only "punishment" they'd institute would be removing your vouching privileges. Not making use of the vouch feature out of fear you won't be able to use that feature seems entirely paradoxical. It's similarly paradoxical to say "I value my ability to participate too much to actually participate" unless I've simply missed some part where they say "we'll take away your submitting/commenting ability."
You mean you had it once, but now you lost it?
Is it more of a concern about not being able to flag something when you feel very strongly it should be, the avoidance of feeling reprimanded by a community you care about, or something I haven't thought of?
Edit: At the heart of it my question is essentially the same as masterzora's. Perhaps I phrased it poorly? I don't mean to be dismissive or rude, I'm just trying to understand the psychology behind this.
EDIT (after downvotes): Because the mod team is responsive to discussion, and they would probably have let you know how you'd tripped the filter, and would possibly have reset it for you.
But I do upvote stuff I like and especially stuff I think is being censored by group think... so that could explain why my flag priviliges where never given or taken away before I could notice it..
The system does screw up sometimes. For about six months I couldn't submit articles without running into a "You are submitting too fast--slow down!" message. I don't submit all that often, so it took me a while to realize that it wasn't just responding to multiple recent comment posts and that something was wrong, but when I finally did email support, they said that my account had been flagged by mistake and fixed it promptly.
This "feature" needs to crawl into a hole and die anyway.
Sarcasm? Irony? Or am I really thinking like that?
Depends on my mood and the experiences I had that day with people. ;-)
https://blog.ycombinator.com/meet-the-people-taking-over-hac...
That said, Dan Gackle (pronounced Gackley, like the town in North Dakota: http://www.gacklenorthdakota.com/) should consider adding his full name to his HN profile: https://news.ycombinator.com/user?id=dang
And much as I like the opacity of the Bruno Schulz quote, it probably would be a good idea for him to also mention there that he is the moderator of the site, and thus the likely recipient of those questions emailed to hn@ycombinator.com.
I think a "mod" tag makes sense.
Once the [mod] tag is invoked the communication is to some extent out of band and fully public. It's a potential gasoline pour.
For example, in the past dang has posted some really interesting content, but that should not be distinguished with a [mod] tag. Similarly interesting but non-moderator comments should not be distinguished.
I have mentioned that also in the past as I am sure others have.
Likewise for sama, for pg, for that matter anyone else that is involved in HN or YC. (Some do of course but for the life of me I don't know why sama and pg do not). All not listing info like this does is simply enforce some secret society of HN. Which is funny given how many comments seem to rally against things like that. It's like "let's shun the newbies and put them at a disadvantage vs. established people who know the ropes and the players..". "Isn't it funny when a newbie doesn't know they are speaking to Sam Altman in their reply haha.
I don't think the way it is handled is perfect or ideal, but community is a hard problem.
Do you think we could maybe work towards removing hellbanning/shadowbanning of site features for non-spam users at some point?
Perhaps it would be good to show the "flag" count to the owner of a post, if that owner is not already banned? This might be provide warning to a perceptive user that their comments are considered inappropriate, while not making it obvious to intentional spammers that they need to create a new account.
For the same reasons, it might also be helpful to show the split between up- and down-votes, rather than just the total. Since this is visible only to the original poster, it wouldn't have much impact on the overall interface.
The quality of discussions is what brings me to HN, and IMHO the quality is poor; it's just better than the alternatives. The vast majority of comments aren't worth my time (or anyone else's, but they can speak for themselves). In other words, there is much room for improvement and I hope that is Dan's focus. I happily would accept unfairness, and suffer its slings and arrows myself, for a higher signal-to-noise ratio. I'd happily lose a few good comments in return for of better quality overall (i.e., false-positives are not really a big deal - so what if my good comment occasionally gets voted down or otherwise buried).
By prioritizing quality over fairness HN can best distinguish itself from a million noise-filled alternatives where 'rights' are the priority. Everyone has a right to their opinions, but not to my time.
EDIT: Several edits to explain myself better.
More importantly, it affects the occasional person who makes mostly good comments but got hellbanned 3 years ago for something that probably could have been handled with a simple downvote and a "please don't do this, here's what the guidelines say" instead.
I regularly see high-quality comments, sometimes the highest quality in an entire discussion, that are marked dead because the account is banned.
One difference between regular banning and live banning is that the banned person is kicked out of the community with a regular ban, but not with a live ban. A live ban mitigates the effects of problematic behaviors on the community while still allowing the person behaving problematically to retain their identity and participate in ways that aren't problematic: i.e. people who are live banned can remain part of the community.
Because the live banned person remains within the community, there is an opportunity for the community to recognize unwarranted live bans based on actual exhibited behavior within the context where it is relevant. Anecdotally, I've seen unwarranted live bans lifted in real time.
I've also seen members of the Hacker News community who have benefited from keeping their identity within the community over an extended period of time while continuing to exhibit problematic behaviors. People may look at that situation and see something different, I see live banning as a highly compassionate way to solve the corner cases that need high levels of compassion.
From a practical standpoint doing regular banning "right" with formal notification, explanation, and appeals processes requires a non-trivial moderator time and energy. If most bans are justified, that means all that energy is wasted on accounts that the owner doesn't value and accounts that the owner values solely as a conduit for argument and/or insult.
Even in cases where the person values their account as an identity within the community, formal processes are an escalation. The most likely proximate cause for needing to limit that type of account is the manner in which disagreement is expressed. Creating a context that threatens identity is unlikely to suddenly produce better behavior in regard to disagreement over a sanction already enforced.
Regular banning is confrontational. One of the ways Hacker News differentiates itself from other sites where people type into boxes is by discouraging that very behavior. Lunch at Cafe Hellban comes at some cost to the community, but I believe it's lower than the alternatives.
But the purpose of "live banning" is to fool a user into believing they're part of the community, while trying to keep them from actually participating through subterfuge. If live banning were simply a way to allow "problematic" users to participate, then the subterfuge wouldn't be necessary, and showdead wouldn't be opt-in, it would be opt-out. Banned users are meant to be ignored and forgotten about by default.
>I've also seen members of the Hacker News community who have benefited from keeping their identity within the community over an extended period of time while continuing to exhibit problematic behaviors.
If you're talking about TempleOS - I would argue that actual banning would be healthier for everyone in his case. He is the ur-example of why "live banning" doesn't really work as intended, or why, at the very least, actual banning should also be an option alongside live banning.
Encouraging him to stay and post while banned is what has turned him into a museum exhibit and mascot for this site, and I think the effect on the community of having to justify this as being, somehow, the best of all possible worlds is toxic.
>From a practical standpoint doing regular banning "right" with formal notification, explanation, and appeals processes requires a non-trivial moderator time and energy.
You want moderators to apply more than the minimum possible amount of time and energy to their duties. But given that there is already an informal appeals process for bans involving someone publicly calling out a banned user and having them email the site, applying that process to a form doesn't seem like much greater effort. Sites with far more traffic and far greater hostility manage with entirely volunteer moderators.
>that means all that energy is wasted on accounts that the owner doesn't value and accounts that the owner values solely as a conduit for argument and/or insult
Of course, that assumption only holds true for some banned accounts, not all. Those obviously deserving accounts could be shown a generic banned message, mentioning it can't be appealed. Or show the generic message and a contact email in all cases.
For other banned accounts, it would be unethical to consider appeals to be a waste of effort because it involves work for the moderators. There is no reason the process has to be unreasonably complicated in any case, as long as a banned user knows they've been banned and has some way of contacting staff to discuss this.
>Regular banning is confrontational.
"live" banning can be confrontational as well, a passive-aggressive, 'slowly poisoning someone's food with arsenic kind' of aggression. Dang now warns people repeatedly before banning them, which is confrontational, but also demonstrably more humane than simply dropping them into the oubliette.
Although, of course, new users may not know what "live-banning" is or when it occurs, they may expect a normal ban. But a user who has been warned repeatedly and then live-banned seems more likely to deserve it than one who's banned without warning.
>Creating a context that threatens identity is unlikely to suddenly produce better behavior in regard to disagreement over a sanction already enforced
Except with live bans, if the sanction is enforced in such a way that the user is unaware of it, the reasons behind it, or the necessity for self-correction, obviously this will not result in a change in behavior. To expect someone's reaction to a ban that they're aware of to be the same as their reaction to one they're unaware of is, I think, unreasonable except in extreme cases.
If a person doesn't know that they've been live banned and removal of the stimulus that triggers problematic behavior creates new habits on the site, I think that's a good thing. If a person doesn't know they've been live banned, and simply goes somewhere else to engage in trolling behaviors, I think that's a good thing too. I acknowledge that others do not.
I don't think they get off on it, no, but the purpose of a hellban is to trick a user into thinking they're visible when they're not, so they get frustrated by the lack of response and give up. That's a mean and dishonest thing to do to someone, even if it may arguably be the best way to deal with certain kinds of trolls.
> One difference between regular banning and live banning is that the banned person is kicked out of the community with a regular ban, but not with a live ban.*
Hellbanned users are not part of the community. The entire point is to remove them from the community without them realizing it.
(BTW, what's a "live ban"? That doesn't even make sense. Did we really need a euphemism for shadowban/hellban?)
You need to understand that hellbanning was invented as a mean of dealing with commercial spammers and malicious trolls--people who have no interest in good-faith discussion, only in provoking a reaction; if they receive a normal ban, they'll just start a new account and continue, because they have none of the investment in the community that would make them care about their old account. Hellbanning prevents them from doing that, and in the old days it was often discussed with no small amount of sniggering over the idea of some asshole troll trying and trying to piss people off and wondering why it wasn't working anymore.
A hellban works for those people; it may be the only thing that does. It is an entirely inappropriate tool to use against a user who wants to participate earnestly but is sometimes excessively rude/short-tempered, and my impression is that HN uses it that way often. (Based on the number of times I've seen commenters say "XXXX, you may not realize that you're shadowbanned, which is a shame because you've got a good point." A person who sometimes has good points is someone who should receive informed bans/suspensions, not hellbans.)
Live banning as instituted on HN via showdead=yes, allows the community to see good content, acknowledge it and engage the banned community member constructively. Because regular bans and suspensions are all or nothing, they don't leave space for those things to occur.
I invented or reinvented "Live Banning" in my post to preempt a potential argument vector over terminology. I think it has the advantages that accrue to technical terms: neutrality in particular. "Hellbanning" suggests the deliberate creation of misery that often forms the backbone of internet fora. I supposed "shadowbanning" means something slightly different, and I may known what it was at one time but those brain cells are apparently repurposed. Anyway, "live banning" has the advantage of not necessarily connoting snickering among its side effects and it takes indicting HN's current mechanisms based on bad behavior by moderators that happened elsewhere on the internet off the table when deciding on the here and now. I fully confess to deliberately reframing HN's moderation mechanisms.
Reframing may be one useful abstraction for thinking about live banning. Live banning is a closure, an execution frame, a continuation, where constructive behaviors are still possible and the effects of negative behaviors on the community are mitigated.
You know this sometimes happens on HN, right? Some people are given many warnings, and then they're banned and told that they're welcome to come back if they stop that behaviour?
The problem, of course, is that sometimes people actually do intend to contribute but are caught in a hellban because they made a single mistake, or possibly even because they said something that bothered a moderator but was completely reasonable for them to have said. Allowing users to "vouch" is a nice counter to this -- if we see a genuine contributor who has been mistakenly flagged as a troll, we the community can fix it.
This is a fantastic change.
And it's fine that trolls create new accounts - their new accounts get banned quickly if they troll in their new account.
And we don't know if there are methods that detect account hopping. I suspect that there are.
It seems to be a very small number, given how many trolls and spammers get hellbanned and keep on posting, compared to how few seem to create new accounts immediately.
I know that it does happen that good comments are banned, but it's not that common.
And this announcement gives us tools to fix it.
The vast majority of comments are addressed to the user to whom they replied. This is the conversational nature of forums, and it's not a sign of poor quality. We're here to discuss things, not post articles to one another.
If it reduces the amount of (justifiable) complaints about banning and allows legitimate content through even in a few cases then it does help to separate signal from noise.
Vouching basically transfers some of the responsibility for overseeing the automated algorithms to the community rather than a central moderator. Doing so will probably affect more comments than you think -- because the frequent comments from long-time posters complaining about how particular posts are affected by the automated algorithm can be replaced by those posters taking direct action.
This alone will improve the S/N ratio of some conversations significantly.
I remember seeing cases when there was a post on HN, and the author of the posted content decided to chime in and accidentally got their account autokilled for spam. Vouching would help resolve such cases without having to explicitly involve moderators.
So far it seems to be helping! I haven't yet seen a case of the above getting fixed by vouching, but those are relatively rare.
Edit: we got one! https://news.ycombinator.com/item?id=10306786. As soon as we see vouches on these we'll mark the account legit so future comments will be ok.
1) How does spinning out YC as its own service affect a) the job posting system and b) the YC class qualifiers in submission titles? / What does "editorial independence" mean in the context of his announcement? YC submissions had a lot of points, true, but I had thought that was attributed to the high YC user base.
2) Will users be able to vouch for [flagged] submissions in addition to [dead] submissions?
3) Due to the vouch system, will the use of banning in general be readdressed, since there is now a way to address false positives/negatives? Shadowbanning was implemented at Reddit as a last resort (that the new CEO wants to remove), and it isn't respectful to the user to not know if they are banned.
EDIT: Reordered to match dang's responses.
(2) Absolutely. Users can vouch for anything that's dead, including [flagged] and [dupe]. I see the notational confusion there; will ask Sam to update the post.
(3) Probably, but I'm not sure I agree with the line you're drawing from vouching to banning. Today's release massively lowers the cost an account of being banned. Instead of having your comments always stay [dead], they're now up for review by your fellow HNers; the community can decide what's good and bad. I'm not sure 'banned' is even the right word for it now— 'under moderation' would be closer.
The reason I say 'probably' above and not simply 'yes' is that there are a ton of issues to consider about it.
If you were to rename the "duped" label, may I suggest "nuked"?
1) native folding of comments, like reddit. I don't understand why this doesn't exist yet, and is the single more important thing to going through the comment list effectively.
2) better algos for upvoting and downvoting comments and threads entirely. This is something that reddit is absolutely great at, the top comments on reddit are generally the best or funniest, depending on the context. The top comments here are generally the first ones posted, or the ones from people already with high karma, and thus either the fast or the already-karma-advantaged get more karma.
This one seems like such an obvious improvement that I've been assuming there's some really principled reason I'm not aware of for why it isn't done.
Accidentally, I've folded comments many times. My main improvement to the extension (if I were to write my own) would be to _remove_ the folding feature and just have new comment highlighting (which I find much more important than folding). Separately, if this is an important feature, why not use an extension that offers it? Is there a reason it needs to be built in?
It doesn't happen that often, but when it does it's annoying scrolling trying to find the next top level comment.
We hackers will put up with really shitty UIs though, as virtually every one of our tools attests. It's like code indentation, we've all had to search for the next root node in an editor without highlighting or collapsing.
Because I only think about it on average once every few days when I encounter an uninteresting large comment tree in an otherwise interesting thread, which I handle by scrolling for a few seconds.
It's very much a "nice to have" rather than a dealbreaker, but for those of us accustomed to the feature as implemented on, e.g., reddit or slashdot, it's a puzzling omission.
Not having folding comments encourages me to read a good random sample of comments (as opposed to only the root comments) which has usually resulted in some fantastic discussion. On Reddit I read root comments and then instinctively collapse the remainder.
That's not to say folding comments wouldn't be nice, I'm just providing some counterarguments for them.
Nevertheless, the conspiracy theories and rebellious "<name in other thread>, looks like you've been hellbanned by the evil mods!" comments are still around, and I suspect that the 'vouch' feature will help kill the last of it. I really hope that it'll work as intended!
(assuming it's all unfounded of course, which I believe it is)
I agree that hopefully vouching will kill the "<name>, you appear to have been hellbanned" comments, though.
Donwvotes on HN are anonymous and silent. This equates to being slapped in the face by someone with a mask on while you're in the middle of a conversation, without knowing why, and not being able to do anything about it. In a real conversation, disagreeing with someone involves actually opening your mouth and talking to that person, and even then is not rewarded with an immediate punishment to the person you're disagreeing with. This is what leads to a discussion, or better yet, a debate. Instead, the only recourse on HN is to change how you talk or to keep quiet.
But at this point I must assume this is an intentional design decision, and I adjust my tone and substance accordingly. But I wouldn't be writing this if I weren't okay with HN's rules.
--
(edit/addendum: this isn't to say there are no discussions or debates on HN... but I hope other's agree there is something uniquely HN about most of them)
There are a few topics I happen to know quite a lot about (e.g. health economics), and I generally enjoy discussing them with people in real life.
I've long since given up on being able to discuss them on HN, however, because it's really demoralizing to write a 1000-word comment with five footnote citations explaining the nuances of the health insurance billing system and labor supply market, only to get downvoted within 30 seconds by people who (presumably) don't like that my comments don't perfectly support their opinions. Rinse and repeat throughout an entire discussion thread.
So instead, on these topics where I know I'll inevitably get downvoted, I just write a single comment, much shorter, and oftentimes without citations. If people respond, I often don't bother to reply. It's not that I dislike debating, or that I dislike debating this topic. But it's a waste of my time to keep writing comments only to get (virtually) slapped in the face every time I hit "submit".
There are other users I've noticed who have followed similar patterns. Which is a shame, because I've learned a lot from them in the past.
The effect of downvotes is to push comments down and fade them away into the background. That should be the effect of filtering low-quality comments, but why should that be the effect of disagreement?
Better, if you can mentally arrange it, is to treat some small number of downvotes as the price of admission rather than a personal attack, and to continue the conversation with those who are listening. The hard part is figuring out which downvotes are actually useful feedback, and which should be treated as random noise.
This wasn't my point above, though. My point was that I like chimeracoder's long detailed comments, and would like him to keep writing them, regardless of the behavior he observes in others.
People could then compare their behavior against the others and one could define a "too often" in terms of std-dev from the mean or a quantile, both within a specific time window (say last 365 days).
In this way one has at least a candidate set of abusers (for the "HN Mods") and everyone could compare his downvoting behavior and adjust.
Additionally, since it's a simple statistic, it should be easy to implement.
Ratios of vote/downvote per user would probably also mean something.
Finally, I'd prefer to have these statistics private for each user, so there is no way groupthink could get a hold.
In real life, I would never throw a tomato at someone I overheard saying something I disagree with to someone else -- but that is HN. By design, I assume, but still.
And sometimes it means deleting the comment. Not because I'm afraid of being banned, but because it's not very good. I've found that deleting not very good comments is good practice. That just saying the stupid thing that I feel compelled to say is enough. I don't need someone to actually read it. I'd rather have people read my occasional good comment.
On the flip side. Sometimes I downvote a comment in lieu of arguing. Actually, it's probably a lot more often than sometimes. Those downvotes are in lieu of mostly unproductive xkcd386 "someone is wrong on the internet" comments I might write if downvoting and moving on wasn't an option.
There are things about which it is reasonable to disagree. For example I believe that discussion is more productive than debate and that conversation is far more productive than either because it doesn't require opposing viewpoints as a starting point and doesn't entail the idea of winning or scoring points.
There are plenty of other places I could do that.
The necessary karma threshold for downvoting is already prohibitively high, but submission karma and comment karma seem to be shared.
That means if you get one lucky submission, such as the announcement of the next major release of a popular project, you likely get shot over the threshold relatively fast. If the gain from there isn't modulated (or doesn't require a separate, sufficently high amount of comment karma), that may give people downvoting privileges too early, before they've familiarized themselves enough with HN.
Seconded. I'd also like to see something implemented to control the contrast on downvoted/dead entries.
Either collapse them ala reddit or give me the option to disable the low contrast view.
In case anyone is wondering, 'flag' and 'vouch' links don't appear beside comments in threads. You have to go to the comment's individual page (linked from its timestamp) to see them.
Good change, by the way. Don't know how many times I've (probably fruitlessly) commented next to a dead comment to point out that they're unknowingly accidentally shadowbanned.
Although HN should have a responsive UI the way Lobster does it: https://lobste.rs/
Please add the ability to undo/change an upvote or downvote. It's very easy to accidentally vote on something when scrolling on a mobile device.
I'd prefer a bit more separation between up / downvote buttons.
I'd prefer the text to be a bit larger, although I accept that some people would hate larger text. Is this the kind of thing that HN could have a setting for? optional style sheets? (Because users can't set their own styles on mobile).
Stylish works great with Firefox on my phone. Doesn't help you if you're using another browser or stuck with a phone which doesn't let you choose your own browser (do those exist?), of course.
I am just curious because I use a 7" tablet frequently and have no problem. It used to crash my browser or tablet or something, some years ago. I changed browsers and it stopped happening. PG couldn't find anything on his end.
It even supports thread collapsing.
I know it is increasingly unfashionable to keep resisting "modern" trends and leave the HN look as it is --maybe see it as a mark of prestige somehow? In any case, please never, ever the change the look. Please.
I must say your nightmares sound rather benign. Mine don't usually end with "haha j/k".
I've hit that thing by accident several times trying to expand open comments because it's the exact same icon most applications use for disclosure triangles, and it's at the exact same part of the post most sites put it-- and there's no way to undo an accidental click of it (as far as I am aware.)
Just a suggestion.
My vote links read '▲sig' for up-vote (signal), and empty line of separation, and then '▼din' for down-vote (noise).
/* remove arrow background images */
/* XXX Bug in chrome/webkit requires padding or margin */
div[title="upvote"],
div[title="downvote"] {
margin: 0px !important;
padding: 0px !important;
padding-right: 1ex !important;
text-align: center !important;
display: inline !important;
white-space: pre;
background: none !important;
background-image: none !important;
-moz-transform: none !important;
-webkit-transform: none !important;
-o-transform: none !important;
-ms-transform: none !important;
transform: none !important;
/* width: 6ex !important; */
/* width: 1ex !important; */
/* margin-right: 10px !important; */
}
/* use unicode text arrows */
div[title="upvote"]:after {
content: '▲sig' !important;
}
div[title="downvote"]:after {
content: '\A\A▼din' !important;
}
/* color unicode arrow text green-up, red-down */
a[id^="up"] {
color: #009900 !important;
}
a[id^="down"] {
color: #990000 !important;
}
/* highlight arrows on hover */
a[id^="up"]:hover {
color: #00ff00 !important;
}
a[id^="down"]:hover {
color: #ff0000 !important;
}http://www.shirky.com/writings/herecomeseverybody/group_enem...
And I'm not making a wry comment on HN moderation, just that this piece never gets old.
Yes, scrolling isn't that big of a deal, but neither is stunning design for the sake of stunning design. (Just my own personal opinion and I genuinely enjoy that you like that premii link; it just isn't for me)
A lot of us can't make out the text when it is so small, so a ton of pinch-zooming is necessary, which is a rough user experience.
(That being said, at least pinch-zooming is an option... nothing bothers me more than when a site actively disabled zooming on mobile with the god-forsaken "user-scalable=no" meta viewport attribute!)
My only hope in saying my comment is that I would have the ability to set the way HN is now as my default mobile viewing version and allow others to view the mobile optimized version.
Once we recognized the distinction between voting and flagging it seemed more natural to have this new feature be on the flagging side, hence vouches. The jury's still out on whether it's a good idea or not, though; if it turns out bad, we might consider reverting to the voting idea.
Regardless of the constant arguments around it, I use upvote/downvote on opinions I agree/disagree with. I'd use flag/vouch for things that I don't/do think should be on HN at all.
e.g., on a Pokemon forum I can strongly disagree with someone who says squirtel is the better than pikachu because water-type has more advantages that electric-type. I might downvote it, maybe. But if someone posts "Pokemon are stupid, you're all wasting your lives you big losers." ...I'd flag that. Conversely, if I posted how electric-type pokemon are the best and I notice a dead-reply to me explaining very clearly a position supporting water-type pokemon as best, I'd vouch for it. While I might not agree with the argument, I can at least vouch that it shouldn't be dead.
How will this affect shadowbanned users, though? Can a shadowbanned user's comment ever be vouched from the grave?
I should add that moderators are still going to intervene sometimes—just hopefully not very often. For one thing, some decisions only we have the data to make. For another, some decisions are based on the values of the site and those are not open to change. But the long-term vision for HN is to make it as self-regulating as possible, and it's been obvious for a long time that the bulk of HN moderation should be done by the community. The only question is what the right mechanisms are to get us there, and vouching is a big experiment in that.
Otherwise, I think both of these announcements are great news and would like to offer congratulations to Y Combinator, Dan and the whole community for keeping this such a useful site and community through all these years, even despite growth and changes to members.
Though they probably will not share the code for rankbanning mechanisms and such, since the code will help out people who might want to go around it.
https://github.com/arclanguage/anarki
Possibly arc-nu?:
https://github.com/arclanguage/arc-nu
At least it appears arc3.1 runs under racket (not sure how long that's been the case, but presumably for a while):
$ racket -v
Welcome to Racket v6.1.
$ wget http://ycombinator.com/arc/arc3.1.tar
$ tar xf arc3.1.tar
$ cd arc3.1/
$ echo admin > arc/admins
# WARNING: running random code from the Internet
# downloaded over an insecure link is not a good idea!
# But a checkout from https://github.com/wting/hackernews.git
# *failed* to run under racket...
$ racket -f as.scm
Use (quit) to quit, (tl) to return here after an interrupt.
arc> (load "news.arc")
nil
arc> (nsv)
rm: cannot remove ‘arc/news/story/*.tmp’: No such file or directory
load items:
ranking stories.
ready to serve port 8080
http://arclanguage.org/item?id=19174
"Ask Arc: How to add a toplabels in news.arc?"
> On Hacker News, the ask and show pages are implemented just like the > front page, but they filter the item list based on title or whether or > not there's a link. > > And yes, that means they won't show items not already loaded into RAM ;)
(defop ask ((p page))
(pagepage ranked-stories* p
[and (askpage-filter _) _]
"ask" "Ask"))
(def askpage-filter (s)
(and (astory s)
(blank s!url)
(~begins (downcase s!title) "show hn")))
https://github.com/wting/hackernews/search?utf8=&q=-filter&t...
Vouching won't change how those topics are treated. And to be honest, I don't know that I mind. HN has a problem discussing sexism in a nontoxic way. I accept the argument that sexism in particular should be a banned HN topic (the way politics is) for the sake of healthy discourse.
Now I think that the conversation on sexism in the tech industry is important, and I think 'healthy' discourse is an illusion used to mask certain political motives, and I have a generally low opinion of HN's moderation policies, implementation, and clarity.
That doesn't mean that banning sexism as a topic would be inconsistent. It would be too consistent: instead it's easier to just penalize quickly commented posts, which happens to frequently apply to topics on sexism.
I just wanted to point out that "interesting fodder for discussion" is not exactly a goal of the moderation team, at least not if it includes "racism or sexism in the tech community."
Is there a way to check if an account is banned?
EDIT: FYI, my karma counter has stopped accumulating points, and popular posts don't get on the front-page (e.g. https://news.ycombinator.com/item?id=10301364 – which has 4 points in 23 minutes – clearly higher than frontpage story 28 (https://news.ycombinator.com/item?id=10300732 – 4 points in 2 hours)
I think the algorithm for story ranking is a bit more complicated than that...
During testing we did notice an unforeseen consequence, but it was a good one. Reviewing vouched comments naturally leads to looking at the history of a banned account, and in a few cases we just unbanned it. Some had been banned for an obvious reason in the past, but recent comments were fine. Others had been banned by mistake, such as by a rogue spam filter. So although this feature applies on a per-comment level, it can have per-account side-effects.
This also raises the question of whether we can take this feature all the way to letting the community manage which accounts are banned (presumably with occasional moderator overrides). That's on our list, though it's too soon to tell if it will work.
Still, I get it, it's a kind of super-vote, reserved for people with a little karma (i.e. some "skin in the game"). Out of curiosity, how'd you guys arrive at 30 for cutoff?
You'd be surprised at how different votes and flags are in effect. The upvoting system is a lot more broken than the flagging system is. People tend to upvote as a reflexive "me like" instead of a reflective "this is interesting". (That's not a criticism—it's simply the chemical reaction of the voting mechanism and human nature.) Flagging is much more reflective in practice. You can think of vouches as an experiment in seeing whether up-flagging can contribute as much value to HN as down-flagging has.
Is it important to the design of HN that the community engage with (i.e., not shun into oblivion) worthy points of view with which it disagrees?
If that is valuable to the design of HN, may we simply accept that voting registers agreement and let another mechanism such as the flagging system be the place to register a comment's value? Perhaps we can display the measure of agreement in a more subtle way than is presently the case and refrain from pushing those comments completely out of the conversation.
I was actually hoping for some read on the two questions I gave because I believe these are important issues, and your comment prompted an opportunity to raise them.
Suppose my click on the down arrow expresses the thought "you're wrong, fuck you". HN is better off when that's the end of it...whatever it was and regardless of if it was in the comment or in me or in both.
I do this for sure. I do a bit of both, using both semantics for different cases.
Thanks for the explanation. I wonder if somebody here would like to take a stab at analyzing user data to see if 30 really is the best karma for "unlocking" these powers.
I think this is a great idea for the community. I applaud this decision because there are quite a few accelerators, but the special part about YC is HN. Being open enough to give it it's own editorial steering wheel is a great indication of goodwill to the community :)
Now that HN is it's own thing, maybe we'll see access(api?) for vote attribution?
Is that even remotely possible? Or, are the implicit privacy issues too much of a burden?
I was thinking it would put rest to the voting rings theory if anyone had access to everyone else's vote history. But, the cost probably isn't worth the benefit.
I'm curious why you'd want this? Typically when voting people don't expect that they're going to be 'named and shamed', or even categorized on any level other than anonymously for stats.
I think HN has gone from "we can expect everyone to pick it up" to needing a few more signposts. That's an interesting post but it passed by on my timezone / holiday / daily lack of caffeine hour.
Eventually we get the knowledge spread around, so it's no biggie. Just an idea.
Edit: of course you could improve the number of reads by changing the headline: "How Megan Fox taught herself Clojure in three weeks, launched a bitcoin healthcare startup that raised $4 million - all from her Apple Air"
Now, hell I would have read that. It's amazing what cat, pipe and the HN headline generator can do.
I've never really understood what those are in the first place, and am curious what they'll be following this statement of editorial independence. Are those advertisements dang does and will continue to pick and choose? Not that I really care, but they've always seemed a bit out of place.
Other than that, I think all live posts are open to comments.
Indeed the nice thing about the current arrangement is that the YC startup job ads are completely up-front, always work the same way, and have a fixed effect on the front page. So it doesn't affect editorial judgment at all. (Except that we try to take linkbait out of the titles and I'm surprised that no YC startup has asked us about that yet.)
I qualified that with "immediate" because an independent HN focused on intellectual curiosity is what's best for YC's business. That's the global optimum. It would be foolish to trade it away for any local one (e.g. handling a particular YC-related story differently). What makes HN valuable to YC is optimizing its value to the community. Everyone at YC knows that.
The main benefit of today's announcement is if it helps us make that clearer to others. Understandably, people sometimes take it for granted that since HN is owned by an investment business, it must directly serve that business. Those concerns won't go away, but it will be nice to be able to point to the org structure and say that's why it's set up that way.
The typical solution to that is to give the users a setting page where they can pick between the two contradictory options.
With some careful design, and not overdoing it on options you can please most (but not all) people.
Go @dan, thanks for listening, then doin' something about it.
I guess I can see people flagging posts that are relatively unconstructive more liberally now, things that would maybe just be ignored before. Maybe that's a good thing, I don't know.
I can think of many times where I've unintentionally clicked the downvote button while meaning to upvote, and had no way to "unbury" the good comment I mean to hoist. I'd hate to accidentally flag and not be able to rectify the situation.
Also, a while back I made a little Hacker News logo concept. (https://news.ycombinator.com/item?id=4731140) If you're thinking responsive redesign work, just hollar. ;-)
I like it.
Mix a random story from the New page into the front page on each page load.
In my profile, similar to [show dead] - let me select the number of new posts to randomly show...
And another idea: always show posts with word [xyz] in title...
Edit: and finally; collapsible threads please
(I've posted about this a few times; let me see if I can find the links.)
Edit: I found some of them, in addition to the ones scott_s mentioned downthread.
https://hn.algolia.com/?sort=byDate&prefix=false&page=0&date...
This would help people that simply don't take the time to click the "new" button often enough - people like me - to contribute to the initial selection much more often.
The simplest option would seem to tag these mixins somehow to make them obvious and, perhaps, allow opting out of this altogether in the account settings.
I'd urge you to try it again, but on a large scale. I know I personally would tolerate one questionable entry on the FP if it frees me from the need to do the /new duty.
____
For article links, it can be tracked with navigator.sendBeacon for recent browsers with JavaScript enabled, and a server-side redirect otherwise.
There is some risk of promoting click-bait this way though, so might need to be careful and provide an "undo upvote" or downvote button, and perhaps find some click-bait-related signals to hedge against (maybe time until next action on HN, titles with lots of generic words, similarity to Buzzfeed headlines, etc.).
I have noticed there appear to be artificially-bumped posts from time-to-time that relate to YC companies. Is HN providing extra clout to YC companies' PR above and beyond organic community moderation, and if so, will HN's extra autonomy in any way change that?
Based on everything we know about HN it's unlikely that such threads would be any good. HN discussions are good when there's a specific story or idea to discuss.
I don't have the data, but probably most programing languages have a set of users that is very interested in them. IIRC this caused a few times a false positive with the voting ring detector a few years ago.
Probably, YC has a mix off many happy alumni here and many fans, so any story about YC get many upvotes fast, and that may look like an artificial bump.
I always assumed not having a mobile style sheet was intentional to slow down eternal September, but if sama is asking for it then that can't be the case, right?
If Hacker News was as adamant about stopping the Eternal September effect as some people believe, they would demand references and a CV before letting anyone join, or else be invite-only like lobste.rs. Not having a mobile stylesheet doesn't really seem like it would help in that regard.
Cynicism and incivility are far more corrosive to this community than the Eternal September effect could be, anyway.
That would not be HN and would be throwing the baby out with the bathwater. I would imagine most users have used a throwaway acct to make some comment without being brutalized, and some of the best comments come from new users.
I'm not sure what your argument is here. Who else would complain? The people who have been driven away?
That said, I wouldn't mind a better mobile site. If we want to be elitist about it, just make it available for users with > 10 karma. ;)
I don't know Michael, but my impression from his writing is that if one could pick and choose just his best comments, he'd be a valuable asset to the site. But if it's "all or nothing", I'm not sure.
I am getting to know Dan a bit, and I feel confident that his choice will be based on what he feels best for the site and the users as a whole, and that he will neither accept nor reverse previous decisions simply due to failure to reconsider.
Second --and this could be quite complex-- let me, the reader, decide who and what I want to read. In other words, let automatic moderation have a light touch and let me, as a reasonably intelligent adult, decide what I want to see.
For example, an obligatory "tech"/"non-tech" tag could allow readers to choose not to see anything that isn't tech related.
I know, tag hell could be horrible. Yet, there probably exists a reasonable set of tags that could allow users to moderate their own feeds rather than having to worry about central moderation.
I could see new posts being tagged "non-tech" by default unless the poster chooses another tag. Tags could include "art", "politics", "religion", "economics", etc.
I'd keep them to a very short set, perhaps five or ten. No more. Too many tags and HN becomes something it isn't.
Within the concept of allowing the reader to shape their own feed there's the idea of "private hell banning" if you will. In other words, if I don't want to see posts from specific people that should be my prerogative, perhaps others are OK with the posts while I am not. Let me mute someone personally rather than centrally muting them as if the entire community thought exactly alike.
I know running communities is extremely hard work. I have in the past and have zero interest in doing it again. HN has really high S/N and that's why I like to read it every day. I've been guilty of adding to the noise here and there. I'm only human. I have tried to self correct when that's happened, mostly by tying to stay away from non-tech threads.
Perhaps you mean that people should stop posting political articles l? (I agree, these are usually hopeless).