Hacker News new | past | comments | ask | show | jobs | submit login
A List of Hacker News's Undocumented Features and Behaviors (2018) (github.com/minimaxir)
673 points by billme on June 6, 2020 | hide | past | favorite | 266 comments

To me HNs biggest feature is the lack of features and the lack of 'innovation', or rather redesigns for the sake of redesigning.

Keep up the fantastic moderation and the wonderful lack of innovation, HN people!

Completely agreed. The site:

1) loads instantly

2) is mostly plaintext.

Reddit, for what it's worth, provides about the same features but on an absolutely heavyweight site. Especially the new (redesigned) SPA reddit

SPA is one of the worst things to happen to the web. There are a small number of instances where it makes sense but it just ruins sites like Reddit, especially on a mobile device.

I've had this conversation a lot lately with a friend of mine who is a React developer. And I recall a thread on HN where someone broke it into a dichotomy.

There are those who believe strongly in the web as interactive documents. I fall largely into this camp. I want small pages with a minuscule amount of JS.

On the other hand, there are those now using the web as a distribution platform. My favorite example is Figma. Figma is my favorite vector graphics program of all time. Compare it to something like Illustrator, but with file syncing, collaboration and updates abstracted. I don't have to expose myself to the horror of Adobe bloat.

So I must disagree. SPA and the rise of JS frameworks is not one of the worst things to happen to the web. Look at what Facebook has become. It's an application, not a website of interactive documents. Or see what GitHub is doing. They are soon to integrate a whole IDE into the browser with Codespaces. I'm bullish on WASM as well. I envision a future where all software could be distributed via the browser.

The problem I think is that too many companies have become overly ambitious and bought into a lot of unnecessary tech. Reddit is a great example, and they are also guilty of pushing their native app incessantly.

The problem isn't the technology. It's the misuse of the technology.

To be completely fair, Reddit is an absolutely terrible SPA to begin with. The infinite scrolling is very poorly implemented (if you click a link and hit the back button, you end up at the top of the infinite scroll page and have to wait for it to load everything below again), the memory usage is higher than any other app on my computer, and collapsing comment threads takes forever.

The Apollo app solves every technical problem with Reddit, so I just recommend that people use that instead of dealing with the official web app. Luckily, they haven't gotten rid of old.reddit.com.

When do you think SPA’s are the best fit?

SPAs are the best we can do for... applications. Online spreadsheet? SPA. Document editor? SPA. Image editor? SPA. Visual programming environment? SPA. IDE? SPA. Facebook makes sense as an SPA too - it's basically a rich widget "dashboard" app with some content creation applications inside of it. Building applications for the web is still not the greatest experience because you're fundamentally building an application over hacks built to expand an abstraction built for hypertext content (the DOM). But, honestly modern JavaScript frameworks, TypeScript, and transpiled environments like Elm all do an admirable job working with what they have.

On the flip side, SPAs are terrible for hypertext content, because existing web technology was literally built for that. Why should a blog be an application? The content creation side, sure, maybe, but viewing a blog? That's literally what HTML was made for. A table of contents full of links like the HN or Reddit homepage? That's pretty much hypertext 101.

How does that compare with https://old.reddit.com ?

i.reddit.com is way more basic than old.reddit.com. It's intended for ancient cell phones or the slowest of slow connections. It's actually pretty impressive that it works at all these days.

Fundamentally all reddit is is some comments and upvotes in a database combined with a ranking algorithm of some kind. Other than scale, the actual core functionality is exceedingly basic, so I hope they maintain these old versions (mainly old.reddit) for a long, long time.

i.reddit.com has infinite scrolling, so it's worse.

You can also append .compact to any url for a simple mobile- friendly version of reddit. I refuse to use the app and this works great for me.

Seconding the Compact version - at one point the site upgrade / redesign reached the point where it was unusable in my browser. Compact saved the day, and makes the experience much smoother to a fellow app-refuser.

The thing that keeps me feeling warm and fuzzy is that they dont redirect any links. It's amazing how special I feel when copy/pasting hyperlinks works like it did back in the '90s.

HN is the one thing I can access on my phone when my data plan goes out.

Ha! I also use it to test my network connection. If HN loads slowly I know it’s the network!

I can't stand the new reddit I'm always on the old.reddit

They are brilliant at managing change, which is why people keep saying "it doesn't change." No, they don't do redesigns for the sake of redesigning, but (two examples off the top of my head) they added a "past" section and they added a character count for titles when you are submitting.

Their motto is "move slow and preserve things." They do change, but not stupidly and they do so with a great deal of respect for what is currently working, which is something many, many, many people in the world utterly fail at it. It is the norm to take for granted the things that are going right, then act all surprised when changes result in throwing the baby out with the bath water.

It's enormously hard to make thoughtful tweaks of the sort the happen here, that get researched and thought through over very long periods of time and go largely unnoticed by the collective subconscious of the site. It's the sort of the thing the world needs more of, but also the world needs more commentary on "Well, we did make this, that and the other tweaks, while leaving most stuff intact." because it does tend to go unnoticed and it does tend to get characterized as "never changing."

> They do change, but not stupidly and they do so with a great deal of respect for what is currently working, which is something many, many, many people in the world utterly fail at it.

An excellent design motto. Respect what works!

They added a lang tag recently, too. There’s small changes that go in all the time.

I do wish there were one exception to this: as a way to nicely format quotes of other comments, the way Reddit does when you prefix a line with > (or multiple > for nested quotes). I actually think totally unformatted quotes are fine or even preferable:

> To me HNs biggest feature is the lack of features and the lack of 'innovation', or rather redesigns for the sake of redesigning.

But so many people abuse the code formatting for these

    To me HNs biggest feature is the lack of features and the lack of 'innovation', or rather redesigns for the sake of redesigning.
I probably see it more often than I see actual code, and it's infuriating to read on mobile. People posting this way don't see the problem themselves (presumably they're posting from desktop), so it's just going to carry on persisting regardless of how "wrong" they are.

At the very least, an instant preview or enforced separate preview-before-posting page would help a lot, especially if either it was narrow enough to represent mobile, or had a warning below not to abuse code snippets that way if there's a detected code snippet. (A preview would help anyway because the reply edit box is monospace, which is really hard to read. I often end up posting an initial draft immediately and using the edit page to fix it up.)

I do wish it had some styling changes on mobile though - hitting vote buttons, reply links, etc. is rather difficult because of how small everything is.

And reading code blocks is the worst. Especially if they are in a far down child reply.

Agreed. On mobile to collapse a comment thread I have to zoom in and hit [-]

Maybe greasemonkey style plugin on mobile could fix it up ?

For sure. This helps to keep it part of the old/weird web like the related articles that get much attention here. I appreciate it.

Its funny, because it almost makes the few of them even more enticing. I didn’t realize that you could change the colour of the top bar past a certain karma, but now that I do there is literally nothing I want more than to get there

I can almost instantly load a hn page, but you go somewhere else similar like reddit? 8 10 seconds of rendering.

The greatest sin in building or maintaining anything is fixing something that ain't broken.

There is one feature I would like to see: NNTP. Failing that, a strict chronological mode.

I wish there was a way of fixing posts that accidentally indented their text 2 spaces.

For me, the biggest feature is the lack of ads. Well except for the job postings, which I strongly suspect are unlabeled ads.

They're job ads from YC companies, as mentioned in the FAQ: https://news.ycombinator.com/newsfaq.html

You're joking, right? On tech topics, every third comment is somebody pointing to their company. Sometimes they mention it and sometimes not but it's still advertisement. Not even talking about people spamming their site.

We have ads here, they just have to be carefully crafted infomercials with a high payoff for the advertisee, enough that it could not be an ad ...

Basically, "Mission. Fucking. Accomplished." - https://xkcd.com/810/.

Often there is too much censoring going on but at least they appear to be a bit more transparent about it now. The one feature that I would like to see is to be able to collapse all comments by default (and read only top level comments).

And often, comments critical of HN like this one first get many upvotes and then get sent back to a neutral 1... which is what just happened

It happened because some users upvoted it and others downvoted it. What did you think was happening?

I was even asked to change my nickname once, I think it was by you. The nickname was nothing bad, you just didn't like it.

That doesn't sound like us, though of course it's always possible that we made a mistake. I've made hundreds of thousands of moderation decisions and tens of thousands of posts. No doubt thousands of those, at least, were wrong.

If you're going to make a claim like this, you should provide a link so that readers can make up their own minds. When such complaints show up linklessly, it's usually because there's more to the story that the complaint is omitting (https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...).

> If you're going to make a claim like this, you should provide a link so that readers can make up their own minds.

fair enough: https://i.imgur.com/Iu7CLxN.png or https://news.ycombinator.com/item?id=16834895

I like Hacker News for the lack of political content. I feel overwhelmed by the amount politics is bleeding into all aspects of my life as of late.

I would like to think that the same type of decision would have been made had your username been from the opposing political party.

> I would like to think that the same type of decision would have been made had your username been from the opposing political party.

I find that hard to believe, but it is possible.


Do you have any reason to believe dang would allow democratic flamebait nicks?

You should state the reason if you have one. Otherwise, you are just slinging a baseless insult.

this was not meant as an insult... this forum is just very democrat, so it would not surprise me if it had a little bias

To be specific: to your knowledge dan has never allowed any troll bait political nicks of any kind.

Additionally: you have a feeling he might allow democratic troll nicks. the strength of that feeling ranges from not surprised if to hard to believe not depending on the day.

Your evidence: claim that the forum is very democratic

It's rare for people to actually supply a link in response to such a reply, so kudos to you!

The description ("The nickname was nothing bad, you just didn't like it") was false though. 'lovetrump' was a trollish username, and we don't allow those on HN because they effectively troll every thread they post to, which is bad.

This is standard practice: https://hn.algolia.com/?sort=byDate&dateRange=all&type=comme.... By the way, we can change usernames now, so the problem is easier to solve.

I don't believe you would choose a username like that (and to a lesser extent, your current one) for any other purpose than to explicitly provoke heated arguments. There are plenty of places on the web where you can go for those, I am happy that dang et al are making an effort to minimize them.

He did say please

I’ll add one to this list, I’m sorry if it comes off as complaining, but at least it’s topical.

There is more than one karma tracking algorithm that can be activated for a given account. That is to say, a downvote is not always a downvote, and an upvote is not always an upvote, and the point score of a comment is not always exactly equal to the number of up and downvotes.

Accounts that are flagged for posting flame-baiting or ideological comments can be switched to an alternate voting mode where votes are not counted the same way. This may mean that any manual downvotes are given greater weight, or upvotes are underweighted, or downvoting is automatically applied after some time providing a type of downward gravity which must be overcome.

I don’t know the precise algorithm. It’s complicated by the fact that I’ve been getting auto-downvoted by bots. But due to some overly combative COVID related posts my account is in this current state. I’ve found that even researched technical comments of mine will inevitably end up at -1 karma, or struggle to stay above 0.

After reaching out to dang about bot-downvoting Daniel was nice enough to look into it and confirmed my account was getting bot-downvoted but also explained that my account had been flagged and made some suggestions on posts that crossed the line. I’ve had a long and mostly enjoyable relationship with HN so hopefully I’ll be out of purgatory soon.

To be clear I have no interest in debating whether the feature was misapplied in my own personal case, but rather just it’s abstract technical merits make for great meta-discussion of moderation techniques for social media boards.

I was going to mention the same thing. This happens on Reddit too.

If your posts are unpopular for any reason you're automatically penalized. Doesn't matter if you're right or wrong, you're penalized for posting anything that people disagree with or don't want to hear.

That's why sites like Reddit and HN will always be echo chambers. Dissenting voices are automatically silenced. Not 100% of the time, but often enough that most will probably never waste their time posting.

I find that Reddit is generally much worse about this kind of thing; perhaps it is the culture or maybe it is the fact that votes are public. If you say something people don’t like, they’ll quickly pile on you. For some reason people there really like going with the flow, and you can’t even reply to clarify without them coming after you. I have found it much less likely that this happens on Hacker News, and people are generally more willing to listen to a comment regardless of how others felt about it.

I don't think it's just reddit, I believe these systems are prone to triggering some primitive human instincts towards group interaction. I can think of a few plausible explanations for a discrepancy in outcomes across HN and reddit. Perhaps hidden scores or the per-comment floor HN uses suppresses it. Perhaps HN attracts a particular sort of personality while reddit attracts a more representative slice of humanity. Maybe reddit is harder to moderate, has worse moderator tools, worse mods, or just too many people. I'm not sure what the answer is, but one way or the other I consider these sort of systems to be failed experiments.

> Researchers from Hebrew University, NYU, and MIT explored herd mentality in online spaces, specifically in the context of "digitized, aggregated opinions".[4] Online comments were given an initial positive or negative vote (up or down) on an undisclosed website over five months.[5] The control group comments were left alone. The researchers found that "the first person reading the comment was 32 percent more likely to give it an up vote if it had been already given a fake positive score".[5] Over the five months, comments artificially rated positively showed a 25% higher average score than the control group, with the initial negative vote ending up with no statistical significance in comparison to the control group.[4] The researchers found that "prior ratings created significant bias in individual rating behavior, and positive and negative social influences created asymmetric herding effects".[4]

> “That is a significant change”, Dr. Aral, one of the researchers involved in the experiment, stated. “We saw how these very small signals of social influence snowballed into behaviors like herding.”[5]


Why would it have to be a bot? If you annoyed someone badly enough, there’s nothing stopping them from camping your comments page and manually downvoting your comments as they appear.

Agree, I personally feel this is a major issue, in fact recently posted twice about it:

Ask HN: HN user ghosted 5 years, why?


Ask HN: Dark Patterns on HN


love dang's response. He sounds like a cop.

It's called being shadow banned. All your comments are basically auto-flagged. People can see there is a comment there but they have to go into your profile to read it. It's happened to me before, but that's because I would make a bunch of shitposts and thankfully they don't put up with that stuff here. Reading your post makes it seem like you know exactly why you were shadow banned (being overtly combative on HN) no matter the content. Make a new account and act civilized and all will be well. You can also ask nicely to be un-shadow banned.

This is generally referred to as shadow banning. When the user is unaware their account is being restricted due to content rules or bot like activities.

technically, aren't you breaking the rules by talking the voting system meta? It's so weird, they ask you not to talk about voting, but then have a feature where you get auto downvoted. How are you not supposed to talk about that?! It makes you feel insane, like someone is following you around.

Imagine someone was following you around IRL delegitimizing your experience. Telling people not to believe you. That your opinions are worth less than others. Only the other people can't see that person, they absolutely believe that person, and people think you're crazy for thinking that it's happening.

I think there are some "features" missing from this list. I seem to recall that Hacker News will transparently remove some characters from titles (such as emoji and exclamation points), which seems like a bad feature to me, and one that people should be aware of.

It will remove many emoji from comments as well, although not all of them. Not sure what the criteria is for that, either.

Not just emoji either e.g. the box-drawing characters (https://en.wikipedia.org/wiki/Box-drawing_character) also get stripped.

OTOH domino or mahjong tiles (🁓, 🀕), playing cards (🃅) or musical symbols (𝅗𝅥) are allowed. So are arrows (↸) including supplementals (⤪), number forms (ↈ), superscripts and subscripts (⁴₂), "miscelllaneous technical" (⎋, ⏱) or geometric symbols (◉), however misc symbols (https://en.wikipedia.org/wiki/Miscellaneous_Symbols) is a no.

So yeah, pretty arbitrary. I guess pg likes playing domino but hates chess.

Also funny note: we have access to musical symbols but not notes, because the notes are in the misc. symbols block. Unless you want to use byzantine (𝀶) or ancient greek (𝈙) musical notation then it's OK.

On FF on (older) Android I get the symbols here:

>So are arrows (↸) including supplementals (⤪), number forms (ↈ), superscripts and subscripts (⁴₂), "miscelllaneous technical" (⎋, ⏱) or geometric symbols (◉), however misc symbols (https://en.wikipedia.org/wiki/Miscellaneous_Symbols) is a no. //

But not the others, they're just get boxes; except the boxes which are missing completely (ironic).

IIRC, we had access to UTF font formats (using bold, outlaid, double, open, etc., letter forms) and to Zalgo, Z̸̠̽a̷͍̟̱͔͛͘̚ĺ̸͎̌̄̌g̷͓͈̗̓͌̏̉o̴̢̺̹̕ ?

> But not the others, they're just get boxes; except the boxes which are missing completely (ironic).

That's a different issue, the glyphs are missing from your machine's fonts rather than HN stripping them.

Yeah, I wanted to use U+266A once and it didn't let me, but random emoji gets through: https://news.ycombinator.com/item?id=23028433

Box-drawing characters are allowed. I made a comment[1] a while ago that used them to make a table. Unless they're only allowed in monospace blocks?

1. https://news.ycombinator.com/item?id=23139517

Interesting! It looks like the box drawing are allowed but the block elements are not, and I just used the box elements in my test assuming they'd behave the same!

Sorry for having been misleading.

That’s a shame, because I could see legit and serious reasons why someone would want emoji in comments. For instance, what if someone was trying to illustrate a point by “drawing” a diagram of a network, using different emoji to represent different types of nodes.

As someone who despises emojis, I am more than happy that they are removed. There are other ways we could show the hypothetical example.

Agree. It's almost as if using vocabulary is wrong. I've been watching how our conversations deteriorate by gratuitous use of emoji. It seems that some seem to think that emoji can cut through all language barriers, because they're universally understood. Unfortunately, the effect I see is newspeak to the detriment of all, as we lose the nuances of language itself, and indeed we become more 'visual only' in our language.

ℑ 𝔴𝔬𝔫𝔡𝔢𝔯 𝔥𝔬𝔴 𝔦𝔱 𝔡𝔢𝔠𝔦𝔡𝔢𝔰 𝔴𝔥𝔦𝔠𝔥 𝔬𝔫𝔢𝔰 𝔱𝔬 𝔯𝔢𝔪𝔬𝔳𝔢?

ʎɹɐɹʇᴉqɹɐ ʎʇʇǝɹd sɯǝǝs

C̸̡̙̹̠̪͊̽̉̃̆̎͝ͅó̶̧͇̳̝̇̓̉̇̋ṵ̷͊̃̓̏͊̌̉͊́̈́͐l̴̛̘͓̪͙̣̩̼͙̂̌͂͛̄ḑ̷̗̫̲̣̺̞̣̖̎̊ ̴̢̤̩̭͇͈͚̜̣̯͚̖̞̪̿̈́̃̌͌̀̿̃͝͠b̸̨̢̧̢͙͙̯͕̼̩̼͚̹̘̟͆́̈́̇̈́̆̅̈̇̾͘̚͝ḙ̴̭̾͒̈͒͂̌̍̈̂͗́ ̴̛̝̲͈͍͉̲͈̗̗̪͖͌͆̓̐̽̊̌͂͛̚͘͘͜͝ẅ̷̧̛͚̜̳̬͙̥̦͈̟̺́̉ö̶̧̠̱̭̻̱̓̾ȓ̷͔͍͙̠̳̬͍̻̱̙̭̖̈́͊͆͛̅̕͜s̸͚̳̥̬̖̲̯͉͈͇̓͊̄͒̒̕ḙ̷̫͔͉͚̻̗̆̋͐̾͊͂̒͛͘͝,̴͕̤̹͙̈̎̊̽̈́̉͘ ̷̨̛̤͈̲̥̙̟͙͍̉̆͂̽́̆̈́́̓̀͗̕͘̚ͅͅF̷̠͕̳̬͚̾̎͌̅̓͌̅͝N̸̡̧͔̰̰͎̝͍̪͇͈̠̻̐̏O̶̢͓͙̜͓̤̰̟͚̓̀̈́͛͗́͘͝R̷̛̞͕͙̘̋͛̒͆̽̈́̊͝͝͠Ḏ̵̨̠̔̀́̌̔ ̷̡̯̹͕͇̗̺̗̰̹̑͂̈́͜͠.̵̢̞̈́͂̈́̄.̸̧̹̫̞͔̣̦̼͙̃̏͒́̽͘ ̶̲̹̺̠̠̗͑͋̾̈́̈́̒̾̒̏̈́̈́̕͠͠͝

Unicode penis?

𓀐𓂸 edit: yeah hieroglyphics are allowed apparently

Get thee behind me, Zalgo!

On some platforms, the emoji is so small that is difficult to determine what it is, let alone what it's supposed to represent.

One that comes to mind is the "thrangrycat" vulnerability. I say "thrangrycat" but the researches insist on calling it . I've taken liberty of shortening the excerpt to the more reasonable reasons, as the answer got more nonsensical as it went on.

>How do you describe the meaning of this vulnerability name?

>We chose to communicate through a visual representation of symbols, rather than “words.” Naming vulnerabilities using emoji sequences instead of other pronounceable natural languages have several advantages. First, emoji sequences are universally understood across nearly all natural languages. Choosing instead of a name rooted in any one language ensures that the technical contents of our research can be discussed democratically and without latent cultural or linguistic bias...

from: https://thrangrycat.com/

Edit: and HackerNews has indeed stripped the emojis used to name this vulnerability.

Another thing that's undocumented is that comments can be "detached" from a post. The comment isn't deleted, but is removed from view, so the participants can continue having a conversation but that thread is no longer discoverable from the post that spawned the discussion.

Is it no longer visible, or just moved to the top level? I've seen plenty of top-level comment threads that start with something missing context, and then is followed by a reply from dang, of the form "We've detached this thread from $url because $reasons".

it just becomes its own top-level comment, and typically is pushed to the bottom.

I wish they had markdown’s list feature


- like

- this?

That's with an extra newline in between, correct? Most places allow you to make lists without the abnormally large line spacing.

* But yes, this is how I

* Make a list workaround

* It is not ideal


Yes, double newlines between. For paragraph-length bullets that makes sense, but for a few words each it is a lot of extra vertical space.

In my opinion, they should implement this by removing the feature that combines two lines without a newline. Most people type entire paragraphs out on a single line (since web browsers word wrap inside <textarea> tags), so I don't really see the point in having this feature anymore.

Oooh I didn't know about /invited. Few stories of very high quality, looks like HN back in the day

/invited is basically the second-chance queue for things that have been personally stamped to be interesting by the moderators, so it's generally pretty good.

It's only a small part of it, the ones that were too old to put in the queue directly, so we emailed repost invites for them instead. It's on my list to publish a more complete page. The types of stories are much the same on the larger list though.

I really wish they'd make RSS feeds for all the different views.

https://hnrss.org/ provides a bunch of them!

> https://news.ycombinator.com/item?id=19212822

Hacker News has so many strange, undocumented things that even such a list is incomplete. I’ve run into entirely new things I didn’t know existed just by using it more, or by happening upon one of ‘dang’s comments…

Don't forget about their API on github.

It's simplistic but you can still use it to write a sh*ttier version of HN.


I just wish they'd open source their We're-Not-Reddit behaviors library.

One thing I've been wanting to do is write a better mobile UI experience for HN using their api. I visit mostly on my phone and I have such a hard time tapping any of the action buttons (upvoting, minimizing a comment thread, etc). I tried a CSS-only mod using Stylish (I primarily use Firefox on Android) but I wanted to do some stuff that the current HTML structure made difficult.

While I appreciate that they finally added some mobile styling, it was very minimal and ignored accessibility best practices for touch target size.

If you use their Firebase API on iOS or Android then you'll need to make a request for each object, whether it's a news post or comment. If you want to present a comment thread, then you need to request each comment individually, so imagine making 400 requests to present a 400 comment thread.

Their Github docs say to use the Firebase SDK, but the mobile SDKs require that you actually have your own Firebase account and you own the Firebase db, you can't enter an arbitrary Firebase URL and have it "just work."

The API has no authentication, which makes it useless for anything other than a reader.

You may find this interesting, and as a bonus it also includes something that you haven't documented: https://news.ycombinator.com/item?id=22788526

That is indeed interesting, but I have a feeling it'll be documented once it launches. :P

Was talking more about the existence of an alpha-tester list, which has some additional features itself :)

Isn't posting, voting etc just done using POST requests?

I'd be interested in a discourse mirror for consumption.

Also needs a section on muting. If too many of your consecutive comments get downvotes you start encountering "you are posting too fast" messages, even if it there was over an hour between comments and several hours since your most recent.

I've always found it kind of ironic (or at least funny) that the code for HN isn't open source.

IMO, a lot of open-source enthusiasts just don't understand the level of hostile/manipulative attention that get directed at sites with lots of high-value traffic, such as HN. There are quite a few users and companies actively trying to get their pages on the front page by any means necessary. It's dangerous to assume that they're not technically sophisticated.

Sites of this size generally need fairly sophisticated rules to withstand these attempts and keep to the original purpose of the site. And they need to be hidden, because if they are known, they will be successfully gamed. There is much more brainpower aimed at manipulating the site than there is in support of keeping it dedicated to its original purpose.

“Security by obscurity”

Hey no knocks, just saying there’s a name for this tactic.

Security by obscurity gets knocked by some people, but it is a valid part of a defense in depth strategy. For example, I run SSH on a weird port, but I still secure SSH just the same as I would if it were on port 22. Running it on a weird port just decreases the amount of spam I have to filter through in the log files.

So-called "security by obscurity" is critically important.

I must disagree emphatically with open-source enthusiasts who believe that "security by obscurity is bad" advice applies to everything. In my opinion, it only applies to a small subset of certain types of software - packages that are meant to be used on extremely large scale, such as web servers, encryption algorithms, and the like. Attempts to apply it to other areas are foolish.

Obscurity is the only possible mechanism for keeping a highly popular link-aggregating site's story ranking reflective of what the community of genuine readers wants to see when under attack by "content promoter" types.

Sure, valid opinion, but I sure ain’t staking my security reputation on the quality of this forum, doubly so since it operates on many hidden and user-hostile rules to get to this debatable quality that I can’t audit.

Frankly I just don’t think they give a damn about the value of open source, at least relative to immediate things, and I respect that.

The original version of HN, or something close to it, has been open source for many years. It's part of the Arc distribution: http://arclanguage.org/.

There have been many changes to both the HN code and the Arc implementation since then, and those are not open source. We've of course thought about open-sourcing them someday, but the problem is that it would be a lot of work to do that, and then a lot of ongoing work to maintain it and respond to requests. Our dev resources are so limited that this is not in the cards for the time being.

> it would be a lot of work to do that, and then a lot of ongoing work to maintain it and respond to requests

You could open source the code, say in a public repository that was read-only, without accepting pull requests or doing any maintenance beyond pushing to the repo whenever it seemed advisable.

Not sure that’s always been true, or at least for sure PG was open about subsets if the code. Generally, there’s reason to open source code, and open sourcing it would likely result in more overhead not less; generally HN goal is to very, very slowly change the code, interface, etc.

Honestly, cloning HN really would not be that hard, cloning dang, that’s another story.

EDIT: Here’s a recent comment from dang on open source HN:


There is an open source fork but it runs on Racket[0], so the Arc documentation isn't always correct, and of course that forum and this one differ in terms of features.


HN is written in Arc that is written in Racket.

I think HN use a slightly modified version of Racket with a few tweaks to be more friendly with the high amount of memory used in HN.

> cloning HN really would not be that hard

Isn't that basically what lobste.rs did?

HN is a very lean operation and there is significant ongoing overhead cost to open sourcing software that often isn't fully appreciated, it isn't free. They may simply be directing limited resources to higher priorities.

I know many cases of software that was not made open source solely because it would require a substantial resource investment as a practical matter.

There are many apps that release their source without supporting it. A simple link to a zip file is a fine distribution model. Git or any other SCM is a common distribution tool for open source, but it isn’t a requirement.

In theory you can open source software by throwing code over the wall with total disregard for anyone else that may be looking at or using that code. Zero overhead, right?

In practice that doesn't actually work because some users will not respect the boundaries you lay out. No matter what you do or say, some significant subset of users will assert or assume the act of open sourcing code places a litany of obligations on the people releasing the code. Furthermore, some of these people will go to great lengths to try to get you to comply with these obligations. At which point you are either doing a lot of extra work you were not planning on doing to make these people happy or you are dealing with a lot of extra and unnecessary personal drama. Either way, it costs you time and energy that you have to account for.

The only way I have ever seen anyone explicitly avoid this overhead was when no one was using their code.

I haven't heard of that happening to anyone releasing code along with their product that is just linking to downloads.

For me I find it funny that HN doesn’t scale. Whenever there’s an unusually active topic, the mods have to scramble to make countermeasures to keep the site up.

That's not true. I wonder where you got that idea.

It is true that the app server runs on a single core and we don't have a lot of performance to spare. But it handles the current levels of active threads reasonably well. The main concern is that if average load goes up significantly we'll be in trouble at some point.

We've got an ongoing major project that will hopefully flatten that curve, but unfortunately it's hard to find time to work on it.

Well, it does runs off of one computer: https://news.ycombinator.com/item?id=22767439

It usually scales well enough to demolish sites it links to sometimes, like Slashdot used to.

This is amazingly useful. I've been on HN for nearly 13 years and I only knew about a fraction of this stuff.

Agree, not sure why after X amount of rep is earned something like this is not featured for just for new users, users who haven’t viewed it, etc.

Ah, there is nothing past 501? I was hoping something else would unlock for user whose karma points are over 9000.

After 10,000 points, there will be cake.

The cake was a lie.

The lie was, however, delicious.

I think the values have changed, I didn't get the ability to downvote until very recently at >1000 karma

No change. It's been 501 for a decade or more.

PG’s phone number used to appear if your rep got high enough, not sure if that’s still a feature.

On his profile page (/user?id=pg) you mean? I don't see it there with 11k rep. It might require higher rep nowadays or indeed be removed.

Yes, that’s PG’s username, or at least the original & official one, but no, that’s not where it would appear.

Here’s example of a prior reference to this feature:


To save others a click, it would appear:

> in the top right corner of the header

65K karma (and #22 on /leaders) and I don't see it. I think that was a joke.

It happens somewhere between ten and eleven digits of karma :)

Not surprising last reference I know of it is years old and PG is busy doing other things now.

Is there a similar list to /leaders but with "active" accounts ranked by age?

/leaders already gets stripped of inactive users by some mechanism that I have not yet figured out.

Maybe it's just growth? The more active users HN has, the more votes a popular comment will get. So active commenters, even if they stay in the same relative ranking, accrue karma faster and faster. The corollary being, if you stop commenting, you lose your position in the ranking with speed proportional to time since your last comment.

No, there are people who "should" be there who aren't. For example, 'pg with 156k karma.

I see... but in any case, those are still sorted by karma rather than by age. I'm just wondering who are the oldest users that are still active ~daily

I agree with most of those policies, but "downranking of tutorials" seems kinda dumb. I could see it making sense for "how to React" drivel that people use for self-promotion, but I've learned a lot from 1-off tutorials I saw on HN first.

Reasoning is HN’s goal, per dang, is to promote substantially new information — unless the tutorial fits this meaning, it’s less of a priority to feature than those posts that do meet it.

Worth noting there’s nothing stopping you from building custom HN searches like this to find tutorials posts, though this would not solve the likelihood of the community posting related comments:



Sure, read all the comments by dang (aka HN’s main mod) here:


I've learned a lot from 1-off tutorials I saw on HN first.

Could that reflect survivor bias (among all tutorials)?

I'd love if I could choose on my user page not to see my HN points total always there next to my name in the top bar. Seeing that often turns me into someone I don't like.

We might implement that as a profile option. One question is where such a feature should fall between minimal (don't show karma next to usernames in top bar) vs. maximal (don't show any point totals or karma about anything). I feel like it might be better to go the whole hog and just have the maximal option. 'nokarma'.

Oh great. Yes, I guess not seeing the voting on one's own comments would be good too. (Looking at the score on my comments is a very small percentage of my HN time, but my HN total score is there every time I return to the main page. Have to learn to ignore it I guess.)

As long as it doesn't mess with the voting system too much! Maybe there'd be many more bad comments if people couldn't see their points total/comment scores.

The maximal version would hide story points as well.

I agree with a more comprehensive 'nokarma' approach. If I don't want to see my point total, I probably also don't care to see up/down vote options, or any other related features.

Set your top bar colour to black.

Haha thanks, I tried it... It 'worked', but unfortunately can't see anything else in the top bar - "ask" etc.

>> If a user has 251 Karma, they can set the color of the top bar in their profile settings. The default is #ff6600. Here's the complete set of colors users have set.

Happy to see that the most #bada55 colour of all is in that list.

Would be nice to see a sorted list and the count for each color. Could even bucket very similar colors so we get a sense for the general HN taste for colors, though I don't know the first thing about bucketing colors.

I use #93a1a1, personally (i.e. $base1 from solarized https://ethanschoonover.com/solarized/)

I use the Dark Reader extension [0], and the standard HN topcolor stood out too much, so I darkened it to #fff0e6.

[0] https://addons.mozilla.org/en-US/firefox/addon/darkreader/

That is a hideous color. I feel like the person who named it was trolling.

thwy didn't name it, its the color with r,g,b = 0xba, 0xda, 0x55.

It's the astronaut's skin color in Kerbal Space Program.

#dadb0d too

Neat. Been here nearly a decade & didn't know that downvotes are capped to -4 for example.

Yes, there’s a lot more upside to comments as a result.

Generally voting signal to noise ratio goes: upvote, no-vote, downvote — allowing more downvote would most like have more negative than positive impact on the community.

Is that a recent change? I'm pretty sure some of my less popular remarks earned double digit negative points.

No, that's been around forever.

Back when comment scores were public via the API, I retrieved the lowest-rating comments for each month. which was always -4: https://docs.google.com/spreadsheets/d/1IfbSDYVBXiHZCuMdHXgp...

Bonus histogram of comment scores calculated in 2014: https://minimaxir.com/img/hn-comments/distribution_comment_p...

It's always been -4 for me. At times it's felt like more when I've doubled down on my unpopular comments with more downvoted comments (and because usually every down vote hurts).

I think the limit probably discourages long winded "all these downvotes are a badge of honor" comments. People who crave downvotes only get four. People hurt by downvotes only get four, too.

Greying out goes with downvotes and send to go in more steps than just 4 ... guess I should RTFGAP.

I've wondered why we can't comment on YC jobs postings on HN. I imagine commenting would be beneficial for questions/answers about the company or position.

I asked this a while back. Answer was roughly that job ads aren't substantive and offer little in terms of actual conversation, and repetitive threads that lack substance primarily attract people with previous grievances. This would lead to the same thread happening every time the company posted an ad.

It makes sense that you wouldn't want to encourage ads becoming actual threads.

Because they are officially sponsored YC related ADs.

Although it is a fact, that doesn't really answer why you can't comment on them.

Generally speaking ADs don’t allow comments. Have an example or reasoning why this would make sense?

Some sites do, reddit and twitter for example. Can be useful

Thanks, wasn’t aware either of them intentionally linked ADs to public comment threads.

Have any notable examples?

Comments are usually disabled on Reddit ads. YouTube banner ads for YouTube videos have comments, and it's always hilarious reading them.

for users: How about "I tried it, I like it" or "this product gave me warts", both things it could be useful to know.

for makers of products HN people like: Comments could draw more attention to the product, more clicks, more sales.

for makers of products HN people don't like: No, can't think of a reason why allowing comments on ads would make sense.

The ads are job ads, not product ads. All of your criticisms don't really apply here. I say this as a person who doesn't like them and rather wishes they weren't here.

Nothing stopping any user from creating posts related to an aspect of the content on any of those ADs, but it’s obvious allowing comments for ADs would potentially cause unneeded overhead for the advertiser; clearly if any YC advertisers want to post on HN, they’re free to do so within the guidelines on HN.

HN doesn’t block negative posts about YC companies.

I'd assume its to avoid meta discussions about the companies. In the "whoishiring" threads you often get people complaining about the companies, which the mods then remove.

Nit: generally these get detached and moved, not removed.

Comment "markdown" syntax with some examples would be a nice addition to this.

While we're on the topic of HN syntax, I really wish they would add proper blockquotes. Quotes in monospace look ugly, especially on mobile, and quotes that simply start with a “>” aren't visually distinct enough, imo. Just indenting a paragraph when it starts with a “>” would probably be enough.

Not allowing quotes drives users to limit their quoting to a relevant part. I observed other places that users have a hard time to limit themselves and context is always there.

It would also be labor saving and booboo preventing if copying formatted text with elided URLs like "https://bla.com/bla/bla/bla/bla/bla/bla/bla/bla/bla/bla/bla/... from postings would copy the full URLs, instead of the elided URLs shown in the text.

Is there some magic way to do that with CSS, replacing the "..." with the rest of the URL when you copy it? (There should be!)

CSS can do that with the "text-overflow: ellipsis" property, but only for fixed-width elements, so it wouldn't really work for links.[1]

Note that if you click the timestamp on the comment, links don't get shortened.

[1] https://css-tricks.com/snippets/css/truncate-string-with-ell...

I think this official page covers all of it: https://news.ycombinator.com/formatdoc

I'm don't think that covers everything! I have used that page when commenting, but I'm sure I've seen formatting not described there.

Really? What kind of special formatting have you seen?

I think that italics,

  code blocks,
paragraph breaks, and https://lin.ks are the only formatting options available. I certainly haven't seen anything else.

I've seen comment citations in the past. I still can't figure out how to get those to work.

> I've seen comment citations in the past. I still can't figure out how to get those to work.

Like this? It's just italics and a > symbol.

Nope, like wiki style citations to external links. Maybe it's just some superscript trickery, but I've seen it from multiple people, so I assumed it's a hidden feature.

I think it's Unicode superscript characters¹. Does that look like what you've seen?

1: https://en.wikipedia.org/wiki/Unicode_subscripts_and_supersc...

Are these the only supported “custom colors” ?


No, you can set any CSS color if you have the minimum karma required. The linked page is just a list of colors that users have actually set.

I really wish any user could set a custom topcolor. I found the default orange hard on my eyes, and I was glad when I could change it.

Mine is #d0c8b5, which is simply a darker version of the page background color. Plain and unobtrusive, and the bit of orange in the "Y" logo sits nicely in the corner.

And you can always use user.css to adjust per site CSS to your preference. I use it for a few sites; there are addons that facilitate it, but I've never bothered.

Not mentioned, but you can be "rate limited", where you can post up to ~ 3 comments per hour (?not sure), else you get a "posting too fast, please slow down" message. It's keeping us trolls at bay

Cool fact! I always assumed this applied to everyone. I've seen this message on and off for at least a few years. Now I'm wondering why/when this was applied to my account. Interesting!

Related top level comment in this thread:


Didn't know that was a thing. Is it a function of your karma?

No, manually flagged

In my case it is 5 posts per 3 hours.

applies to submissions too

"Posts without URLs get penalized." That's strange. Does that include "Ask HN" posts? I would have thought that submissions with no link would be good discussion starters.

A blog post fills the same purpose for starting discussions except it requires a higher level of commitment by the author. Hacker News is not really a micro-blogging platform in the traditional sense, and because other alternatives exist, the higher bar appears to have some logical rationale. It creates a space somewhere between StackOverflow and Reddit for "I want to talk about X" threads. People who really want to talk about X can write a blog and submit it to HN if Reddit isn't good enough.

My experience is the post without URLs, but say 10 upvotes are generally of lower quality than URL related posts with 10 upvotes. Not sure why this is, but I was forced to speculate, URL post intend tend to be by an expert on a topic and while non-URL post tend to be seeking experts; again, just speculation, might be wrong.

“Tell HN” and “Show HN” are a thing, too. I was recommended to use “Propose HN” once, I think. It’s just text. I assume the ask/show links check title prefix.

Could this be why most Ask HN posts are desaturated? I always wondered why that was the case.


The text color is lighter, as if the post had been downvoted.

That's the 'visited link' color, which indicates that you've visited a given post's link. Posts do not change color based on votes. For posts without a link, reading the post requires still visiting the post's link, so it works out as expected from there.

No, the actual colour of the text on text-only posts is super light. I read somewhere a while back that it was supposed to discourage people from using it too much.

Here's an example from the front page: https://news.ycombinator.com/item?id=23438930

Ah, thanks for clarifying.

On my computer, the default color for visited links is also lighter.

I think the minimalistic design is appropriate and efficient mostly.

…with a couple caveats regarding accessibility: the default font is way, way too small, and some colours don't have enough contrast (eg, the “visited link”). Also, a tiny bit more of formatting would make comments more readable while keeping them sober and focused (eg, blockquotes, monospaced inlines, true hyperlinks).

Is there a simple link to get the top stories from the past week? /best seems to only return from the past couple days.

I was always kind of surprised that there's a relatively high karma threshold for downvoting, but seemingly not for upvoting. Seems like it would invite voting rings. I guess there's already other software to detect that though.

Downvotes per item are limited to -4 and the signal-to-noise ratio for negative votes beyond pushing them below the “new comment” boost is of little positive impact in filtering content, in my opinion.

And yes, there a lot of filters in place for upvotes, many of which are intentionally kept secret.

Personally, I don’t use downvotes.

It's easy to engage in anonymous encouraging behavior with upvotes. Anonymous discouraging behavior via downvotes has a higher threshold.

Thanks for posting this. I've been an avid reader of HN for years (with a few modest/minor submissions). I had no idea about a substantial amount of this... I wonder how much is commonly known in the community?

Does anyone know why the top bar sometimes has a thick black line underneath it?

The black line is a symbol of mourning, used when someone deemed significant to the community passes.

minimaxir should add that one! It doesn't look like it's on the list.

Huh, I thought I did add that. Will do!

Can we make feature requests here? Hah.

I would absolutely love it if the comment box was taller than 7 lines on mobile, perhaps just at least when editing a comment.

Here’s a mod’s recent comment on feature requests to give some context on the topic:


If you set the delay profile value to X it’ll let you edit and save a comment for up to X minutes before other users can see it. I use it extensively on mobile and have mine set to 7.

Ooh, that’s a good one because I tend to post then edit, edit, edit, so a little delay would be nice. Thanks for the suggestion!

It took YEARS before they added collapsible threads despite its obvious usefulness and easy implementation. Some people added extensions to do that but HN was entirely unresponsive. I wouldn’t hold my breath for any new features.

I find it interesting that politics and diversity each have their own section on here. I always considered diversity a political topic.

I suspect that it's more about optics than practicality. Not all diversity topics are political, but most diversity topics are probably outside of what the mods believe to be within the purview of HN.

Making it a separate category bin bypasses all of the hemming and hawing over particulars.

The front page way back feature is pretty cool.

One thing that I'd like to know is if there is any kind of penalty if the stories you submit get flagged.


Thanks! Pretty useful. Can anyone supply a good comparison between Hacker News and Reddit?

Stating the obvious, Reddit was in the first YC batch; YC is HN’s parent company. Further, HN was created in part because Reddit’s intend is more general than HN; HN’s intend is to focus on substantially new information that triggers both curiosity & notable dialogues.

Tip: intend is a verb. She intends to win the election. What did you intend? Paul intended not to lose.

Intent is a noun. That is her intent.

Intent is also a verb: She is intent on winning the election.

You wanted "intent" as a noun in both your usages above.

Thanks, agree that’s a typo; unfortunately I am unable to fix it since editing is now locked, but appreciate you taking the time to point it out.

I don't think intent is a verb. It is either a noun or adjective. In the sentence "She is intent on winning the election" the verb would be "is" and "intent" is an adjective that describes her.


His program focused on dinosaurs. (verb) His program is focused on dinosaurs. (adjective describing his program).

You would never say "his program intented on dinosuars" because it's not a verb.


“Moderators will sometimes rescue a post which didn't receive a lot of upvotes and reset the submission time on the post.”

This sounds more like manipulation of content based upon moderator viewpoints or interests. HN has a wide enough audience to hit a good post the first time around.

> HN has a wide enough audience to hit a good post the first time around.

Unfortunately, it is widely agreed that a lot of good quality content goes unnoticed. Which might be due to the sheer number of submissions.

That could be seen in yesterday's post about quality content that goes under.


/leaders is pretty cool. First place got over 300k points, wtf.

Related link, leader list:


Direct link to top users by rep:


(tptacek’s comments are super useful and extremely knowledgeable)

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact