Hacker News Reader

Comment by conesus | Hacker News Reader

I never publicly announced this but I loved Slashdot’s friend/foe system so much that I built it as a cross browser extension for Hacker News. It’s called Hacker Smacker and it’s on GitHub.

https://github.com/samuelclay/hackersmacker

Supports not only friends and foes but also friends of friends and foes of friends. Makes it easy to scan the HN homepage and comment threads and see what’s good. Much like how Slashdot’s friend foe system highlighted the good stuff in threads.

View on HN

Slashdot got a surprisingly large amount of stuff right, amazing considering it was one of the pioneering web-based discussion systems.

- You cannot both moderate and comment on the same article.

- Limited moderation points (too limited on Slashdot arguably, but better than infinite up/down votes).

- They didn't have up/down, but a system of "Interesting", "Informative", "Off-Topic" and a few others. These are the same as up/down votes in the end, but make you classify postings.

- Set threshold to (say) 3 and quickly see only the +3 interesting comments on an article.

- Meta-moderation didn't work well, but was an interesting idea.

When I was younger I thought that Slashdot's moderation was brilliant, and every time that I tried to think up something better I invariably came back to that model. When I first stumbled across Reddit I thought to myself that it couldn't possibly work: too simple, too accessible, too easy to game. And yet over the years I kept reading Reddit and stopped reading Slashdot, and it dawned on me that Reddit's comment rating system miraculously was better than Slashdot's somehow, in a way that I couldn't put my finger on. Somehow their algorithm for knowing which comments out of thousands to place above the fold is basically perfect; even when I click on the "load more comments" links at the leaves of comment chains I nearly always find that the originally loaded comments were exactly where I wanted to stop reading. And of course Reddit still gets plenty of visibly-upvoted bullshit, but the difference is that on Reddit the top response to that comment will be a refutation that is just as upvoted, if not moreso. In contrast, on the rare occasion that I visit Slashdot these days, I wince at all the confidently-asserted yet blatantly wrong comments marked with "5, Informative", while the informed refutations languish invisibly below with a score of 1. Maybe Slashdot's moderation would work better on a more focused site, but in practice it just seems like the people selected for moderation just aren't often enough experts on the topic being discussed to distinguish the signal from the noise.

On Reddit, you get the same thing in certain subreddits. I think it just depends on how you are in relation to the local groupthink.

The "fall of Slashdot" had nothing to do with moderation and everything to do with the fact that articles are chosen centrally. That meant that Slashdot could never expand into other areas outside tech or dynamically follow its users' tastes.

>could never expand into other areas outside tech

Why would anybody want to?

I think it's similar to Netflix replacing stars with thumbs up/down. They realised they didn't need that data, they just needed more dimensions (ie for you to rate more of their library).

Slashdot didn't need to limit voting so severely. And it turns out replies to funny comments are usually funny and interesting comments are usually interesting.

They also never scaled it - +5 was the maximum from when an article might top out at 200 comments or 5000 comments.

> They didn't have up/down, but a system of "Interesting", "Informative", "Off-Topic" and a few others. These are the same as up/down votes in the end, but make you classify postings.

I might be misremembering things, but I seem to recall that it was also possible to assign boost points to these in your user settings - i.e. so that Interesting posts would be treated as +2 rather than +1, for example.

I have wanted HN to have a friend/foe system since I first found the site. I'd also like it to have the ability to make notes, like 'farkies' from Fark.

On Slashdot, you simply add them to a friend or for list. On Fark, you get to flag them in certai colors and then leave yourself a note (other people can't see it but admins probably can) that I use to remind me of why I put them on the list.

I'd like a combination of those two things. A simple friend/foe list with comments that show next to their name. I'd have no use for ignore functionality. Also, I usually use said note to write polite things that help me remember the user.

I think it encourages getting to know the other people and humanizes the pixels on the screen. Both sites have led to my meeting people in real life and making real life friendships. That's easier, for me at least, when I can more easily identify them as individuals and remember them.

Slashdot is where I first read your comments, then you didn't post for a while and I found HN and you post here too!

I didn't create an account on /. since AC was easy and I mostly lurk anyway but I always enjoyed the insights that everyone brought.

I remember either around the time of the revelation of the NSA phone closets there was a guy who used to do SIGINT and he posted a bunch on how to avoid surveillance. I think he had moved to the Philippines, sometimes what happened to him.

Just wanted to say hi and that you seem like an interesting guy!

Thanks! I took the karma hit to say hi to taco, in the thread. Worth it!

I second this. At the very least. some way 'favourite' a HN user, and their name on their posts appears in a different colour. There are so many good commentators here, that keeping track of people is hard.

something like this?

https://imgur.com/a/wRtfp

NB: the icons beside your name are for "model", "traffic", "acquired" and "phd"

Neat!

Funnily enough, every time I see the friend/foe dots/pills on a post my first thought is "that's an old-timer". I'm betting well over 75% of the current Slashdot readers don't even know it exists.

I have a four-digit Slashdot ID and I don't remember friends and foes. Is it after my time, or is it just that I don't remember anything any more?

You just don't remember anything anymore. IIRC, it was implemented around 2001. It was little more than a curiosity though, you may have just bypassed the little colored "pills" on comments, but aside from being angry about having foes, or happy about getting a new friend, nothing was really done with the feature.

I think that friends' posts had an extra point of moderation calculated for your display, though that feature might not have been present in early incarnations.

The point of it was letting you give free pass from the moderation for some people's comments always appearing for you, or hiding other people.

It still exists if you have the original discussion system enabled. I never moved on to the javascript-based one.

I don't really understand the appeal of a friend/foe system. If I don't like a person, I typically try to avoid interacting with them unless/until I find a way to reevaluate them or myself so that I can find something positive, a way to like them. Pretty much the last thing I want is a computer to remind me that I don't like someone. In fact, I wish frequently that I could suppress my own internal feelings of dislike for people. When reading comments on sites like this I try not to read who posted them until I've evaluated the post on its own merit. The times I find myself pre-judging a post when I see who made it, I kick myself and struggle to evaluate its claims as objectively as I otherwise might.

I do understand that not everyone interacts with every site in the same way, and I personally do not approach every online community in the same way. So, I say, to each their own. The use of a friend/foe system seems to impose a bit of centralized structure, however, and I'm not sure how comfortable I am with the idea of enumerating to some website exactly whom I prefer as people, when I'm constantly trying to reevaluate and expand that in myself.

Slashdot is interesting in that it is one of the remaining bastions of (at least superficially) anonymous discussion. There are apparently several strata of users: some don't read anything by anonymous cowards, some don't read posts by their foes, and some only use the site anonymously. It makes me curious about how these different self-selected filter bubbles might give different perceptions of the discussion to different participants. Personally I have great difficulty engaging in public discussion without trying to read and understand as many diverse opinions as possible - I feel as though otherwise I'd be speaking out of ignorance.

EDIT: On occasion (at least speaking for myself) we don't like someone because they remind us of something about ourselves we try not to acknowledge or think about. Blocking their posts en masse can deprive us of an opportunity to explore this and grow.

The whole point of friend/foe system on Slashdot is that foes' comments are less visible, and friends' are more visible. It's basically a way to say "this person tends to say ridiculous things, and I don't care to read more of them" and "this person tends to say interesting things, I'd love to hear what they have to say on other subjects".

So if your approach is to "try to avoid interacting with them", it just automates that.

I used the friends/foes on /., but never thought much about it. Filtered out a number of vocal nutcases.

But wouldn't this sort of create your own cozy echo chamber, where all opinion is streamlined?

I think it would be interestring going the other way around, a bit like some sub Reddits don't show comment score for a day or two, to not cause self reinforcement in moderation.

Maybe even take it a bit further by anonymizing the commenter's name for a limitid time, could be interesting as the comment would then rely fully on content, not reputation or score feedback reinforcement.

It depends on how you use it. If you only friend people who agree with you, then yes, you'll get an echo chamber. If you friend people who say interesting things instead, you won't.

> So if your approach is to "try to avoid interacting with them", it just automates that.

Well, I didn't really mean that I don't read or listen to what they have to say. I just choose not to engage with it.

This is awesome!! Thank you!

Hey - it's a little hard to see the splits like foe of a friend (I'm red/green color blind), what's an easy way to tweak the colors?

EDIT: I unpacked it and tweaked the CSS. Wooo!

Been looking for something like this, as i have been using RES over at Reddit for a similar purpose.

The one thing that the friend/foe system missed IMHO was the ability to tag posters with comments like "suspected oil shill" or "actual Intel chip designer" or stuff like that. That's often more useful than some vague idea that a poster had some especially good or poor posting in the past.

The RES browser extension does this for Reddit (which I suspect you might already know, but others might not). It also sort of does automatic friend/foe for you, as it displays the net number of votes that you've given to each user next to their username, turning bright green for positive numbers and bright red for negative.

The Hacker News Enhancement Suite does this, plus other things.

It's not compatible currently with Hacker Smacker, but fixing that takes one line of tweaking.

Oh, that's nice!

It's not currently compatible with Hacker News Enhancement Suite. Quickest solution I found was to change a line in findCurrentUser(), but just FYI. It would've been a deal-breaker for me.

Looking forward to playing around with this :).

@mercer - There a PR or Fork for that?

I've never done one of those, but if I find time this weekend I might try and do that. If anyone else cares to, or if you want to fix it yourself:

- open the folder for the extension (name of folder can be found in the extensions screen in Chrome)

- edit client.js

- go to line 25 or the function 'findCurrentUser'

- replace with this.me = $('.pagetop .nav-links #my-more-link').text()

fascinating :D

I do have my own system on following/starring PoI in HN, using no extension at all, so I can instantly know who's comment is more useful

no banning though, since valuable nuggets on knowledge can come from anyone

How does that work?

just using Firefox only... Set the color to always use your own colors, then change accordingly

whomever you want to take interest to, will pops up in 'Visited Link' color, the rest just normal URL

obviously works best in less-hyperlink sites like HN, wikipedia quite tolerable

YMMV :D

I read the Slashdot FAQ on the Friend/Foe system, but I still don't understand it. Is the idea just that it lets you filter a comment thread to emphasize comments written by people you're interested in (friends)? Does it hide the comments of foes?

Slashdot's score weighting is configurable. You could even add a positive score offset to foes.

The impression i have is that it gives fores a personal -1 on their comment score. So unless they get some +1s from others, or you have set /. to show everything, they will vanish below the breakpoint.

Note though that anyone can look up who has them flagged as foe on /..

Yes, alter the displayed scoring.

Also based on the 2nd degree information, friends of friends and such.

I would use it behind the scenes to look for bad accounts too. Someone with lots of foes would be a candidate for some silent moderation.

Reply on news.ycombinator.com