I'm trying to read discussion from a thread with 133 comments, on a topic i find interesting, but after a few comments i get overwhelmed. How do you do it?
It took me a while to realize that there is a "prev" and a "next" button next to every comment. It makes navigation much easier when there are long threads. Using [-]/[+] is another option.
If the problem is just the sheer amount of data and not navigation, you can simply bookmark the thread and come back later. The algorithm usually does a good job getting the best comments on top after things have settled down, just read these. Plus, if you forget about the topic and don't come back to it, maybe it wasn't that interesting after all, that's a good filter.
As resident layman and LLM/chatgpt hater, I am offended on behalf of dang. He is not replaceable with a real person; an automated replacement is insulting to even consider.
I didn't say replace. I said replicate. Or, put another way, capture the essence of his moderation style. This could be useful for the rest of the world that's impoverished in dangs.
That's what i ended up doing after i wrote the post. I was reading every parent comment as something different on the same topic and when i finished or wasn't interesting to me i minimized it the [-] button.
I begin by reading the top comment, if it's interesting I'll read its replies, if it's not I'll collapse it's replies and move down until bored.
Most big threads have many replies and far fewer top comments so 133 can become much smaller this way.
I also have no problem quitting as I scroll down. On big threads that have been up for a while, the insight tends to fall off with the number of votes as you keep scrolling.
After getting bored, I begin randomly reading comments as I scroll. I'm surprised at how I find nuggets this way, though mostly it becomes more of the same.
These days I plug the article into https://kulli.sh which aggregates all the comments across HN, Lemmy, Lobsters, Reddit etc. into a single feed which I can quickly filter by source, discussions or order reverse chronologically.
When I've had enough of a particular comment thread or source, I click/tap to minimize it and continue with the next one.
I like having all the comments on an article in one place, not just for the convenience factor, but also because it gives me regular glimpses outside of my bubble and shows me how different communities, including ones I don't necessarily agree with, are discussing the same topics.
The service-side rate limiters for requests made to the various aggregation sources are global across all user requests so it can slow down a little bit sometimes.
I've been thinking of offering dedicated instances for a fixed monthly cost if people are interested, but so far I haven't had any takers and I'm happy enough using the global instance for now.
There is also one I cannot find at the moment which let you click and open each comment thread in the right side of screen, that was very good.
edit: found it: https://hzn.jero.zone
I use a short userscript to “mark as read” comments on hover. They just get a different color so I can hop around and know where to resume. It’s particularly helpful with active threads where new comments appear in the middle.
- The author of the parent comment becomes the highlighted "OP."
I have a bookmarklet that toggles between viewing on HN and my web app.
My app was heavily modified, but original credit goes to https://hackerweb.app/. The original author had an idea to make HN threads more like Reddit threads, which he may have implemented in the mobile app rewrite:
I minimise threads that get more narrow or argumentative than I'm interested in for a particular topic.
Sometimes in a particularly large discussion I'll do that even for top-level comments without replies that I've either read (and won't reply to/not interested in replies to) or skimmed/read enough and decided I won't read.
Help others by voting (both directions) and flagging/vouching where necessary.
I typically use the "next" links (above each post) to jump to each top-level comment, drilling down if appropriate.
Other than that, I try to resist the urge to read everything. Its almost always better to absorb the gist of the conversation than to try to absorb it all.
I skim top -> bottom, read what catches my eye, skip the reply tree to posts I'm not interested in, and when I am not getting value I stop and move out, regardless of reason for not getting value (direction the discussion takes, posting quality, my mood or attention, etc). I almost never revisit later, I don't use any navigation helpers like [-] or prev/next.
If the topic is interesting for me I just wait a day or two so the comment activity subsides and then just collapse each comment I read. Works perfect.
For threads I am interested in but want to comment immediately, I only collapse those comments that I don't find interesting -- and as above, go back a few days later to read the others without constantly having new ones added in real time.
There's the [-] button on the right which hides a subthread. Then go to the next. On particularly huge ones, just go to the next page.
From that, it's like a party. Just go out and talk to someone at random. Past a certain level of discussion, there's going to be repeated conversations on the topic anyway.
It always depends on the community. On Hackernews, for example, I only read the top 5 comments and then check the thread 2-3 times later to see if anything has changed. On reddit, for example, I don't read any comments at all because there are too many bots replying.
Otherwise, I try to keep my own thread when I read the comments. For example, when I read an article about a topic, I try to find an extremely positive point of view from the comments, but also an extremely negative point of view and then link this to my own thoughts.
As already mentioned in the thread, I would like to have an online tool, for example, which reduces comments to a clear point and thus makes reading easier for me.
I sometimes see HN commments that suggest readers prefer nesting. I do not. I dislike indentation; it is distracting. I prefer left justification. I use a text-only browser that presents threads as flat when tables are disabled. I can search through threads of max length using vi mode keystrokes. Much faster than using a graphical browser. Also, I do not use colors so, e.g., "greying out" comments has no effect. All comments appear in the same color textmode font. I decide what comments to read without unsolicited "curation" from voters or flaggers.
No need for hnrss.org. For me, it's not the site that one uses that changes the "user experience", it's the client.
Using a non-graphical client I see all comments in a thread in a single page, all at once. Up the maximum that HN will display at one time. Nothing is hidden.
I’m thinking about making a tool to do this, I haven’t quite gotten to loading a comment thread but I have loaded 250k images and also 400 from Evernote. Next in line is to load some unwieldy discussions, particularly from a site with a horrible comment UI such as Arstechnica.
I’m particularly thinking about how to make sense of badly organized threads from Mastodon and similar things.
With timecode metadata and heatmaps to highlight rapidfire back 'n forth and stuff?
Semantic scoring to rate quip chains Vs. slower, longer considered interactions?
Ways of highlighting high frequency commenters, perhaps rating frequent flyers by degree of interest in what they have to say and weighting threads by "participant rating".
There are a couple of angles, straight up representation of forum | subreddit threads after the conversation has moved on, "live" tracking comments as conversations progress, and moderation of live threads (including swatting | detecting trolls, spambots, griefers, etc) (oh, and retro sweeps looking for tail end spammers and their puppet networks eg: https://news.ycombinator.com/item?id=40800160 (with "ShowDead" on)).
In post analysis there's the whole ball of wax around decrufting chat threads to pull the meat best served to (various) AI, either for training, company | group intelligence, etc. Oh, and fingerprinting across multiple sources looking for common {group | individual} traits
The intrinsic issue with excluding all but those with a current PhD on chat would be the tossing of, say, those with dated PhD in theoretical physics or somesuch.
@dang obviously mods here and likely has a slew of lisp-y scripts | tools for giving various views, @jedberg was about for much of the early evolution of reddit and took part in a lot of discussions about presentation and implemented a few.
There'll also be useful input from any of those who've moderated | admin various largish forums | channels, etc over the years.
I'm interested in fora other than HN but HN is a great target because it is easy to get at through the API. Other fora have the problem of atrocious interfaces such as
or the half-baked "threaded discussions" that exist on Twitter, Mastodon and clones that people abuse in various ways. Those forums need it more.
So far as the PhD I brought that up as one specific example. (Funny I have a physics PhD from a while ago, I still remember a lot about Hamiltonian chaos and its quantum manifestations, can understand something looking at the math from string theory, etc.)
My problem is related to the moderation problem but is different because I don't need to be fair to anybody or be perceived to be fair.
In general I am talking with a friend about ways to make the commenting experience online better that include everything from tools for power users to large-scale modifications of "the system" to change the incentives to change people's behavior. I can demo the first one today, the second is a bit harder.
You’ve got to enter the thread with the expectation that there are idiots posting drivel that you can leap in to correct, obviously.
Seriously, though, reading 133 comments usually takes like two to three minutes total. You need to decide whether you want to spend that much on this site doing that or not. I expect for most people the answer is “no” and I commend them for their self-control.
For topics you actually really care about, you’ll find it a lot easier to read through the comments. You will learn which comments to skip (nobody wants to read a rehash of “does anyone else think all developers are lazy and software is so slow now”) and which ones are actually interesting to read.
I don't bother trying to keep up with everything. I'll follow /newcomments and I have a browser plugin that shows when a thread has new comments but beyond a certain point it isn't worth it. 90% of comments aren't worth reading anyway,
The "Highlight new, unread" future already exists, but you have to email the mods to have them enable it. I've had it on for years. I'm not sure why it's not a GA feature.
I usually/only read comments with replies to get different opinions and go to next sibling comments if child comments aren't interesting to me anymore.
Most discussion systems are inferior to 1990 Usenet, 34 years later.
At this point it is traditional to shout about Usenet not having good moderation. In fact, it is possible for either individual news groups or specific news servers (that only carry a limited number of groups) to have good moderation. It is, of course, expensive: you need to get enough trustworthy people to run it to cover the time and volume.
Then someone will sniff about not wanting to use a TUI, which can be rebutted by pointing at all the GUI newsreaders -- having a standard protocol for talking to a server is quite valuable.
"Isn't Usenet a huge space and bandwidth hog?" Yes, Usenet was a huge space and bandwidth hog -- in 1990, and 2000. If you don't carry the warez groups, it can be adequately handled on a high-end server from 2000 -- which today is equivalent to a Raspberry Pi. (1994: 7 x 2 GB SCSI disks, 1.5Mb/s Internet connection)
HN approximates the activity level and content size of 1 very active Usenet newsgroup.
A mouse with a scroll wheel and all the comments fully expanded on one page and a high-resolution monitor showing lots of lines of text at once.
The page-down button is a distant second. I find using a laptop without a scroll-wheel mouse mildly displeasing. A laptop with a screen that can show a quarter of the amount of text that my desktop PC can do is infuriating.
Are you trying to read long threads on a smartphone or tablet, pawing at the touchscreen with your finger?
I use the HACK app, a hacker news reader app, for android. It has collapsing sections and jump to next top level comment. Between these two functions reading long threads even with deeply nested comments isn't too bad.
Skim for keywords you're interested in, decide from a glance whether or not a comment chain is just two people passive-aggressively fighting semantics (and whether or not those semantics matter to you), scroll to the bottom to see if the dead comments have anything useful to say and rescue them if they say something interesting (albeit not popular)
Don't internalize everything you read here, just a quick dip down for some interesting perspectives, and then you resurface with those endorphins and wait for the next hunt
This is strange to me. Flagging should be reserved for comments that flagrantly break the rules. There's plenty of stuff I disagree with that is formulated in a perfectly polite and serious way. I come here to have my own views challenged as much as anything.
I wonder if people who abuse the [flag] function as a “mega-downvote” should eventually lose the ability to flag. I quite often see comments flagged here that are not rule breaking, they’re just unpopular takes.
It depends on the size of the community in my experience. Old school bulletin board style fitting are great once trolls and manic megaposters are kept under control but if they've too few users they tend to die and if they've too many they become impossible to follow
You mean insane; every sane discussion place looks more like subreddit (or, often enough, is a subreddit).
> Trees devolve into a 1on1 discussion, also other branches can have the same discussion, leading to redundant arguments.
1 on 1 is a feature, as it's never truly 1 on 1 - anyone can jump in instead of just listening to the conversation.
As for other branches having the same discussion, the solution is the opposite of what you're proposing: instead of flattening tree structure into a line, embrace the nature of any discussion (or knowledge accumulation effort) and make it a directed graph. I'm yet to see it done in production (outside 4chan maybe), but that's the right way.
Or, you can just ignore the problem and let people having their redundant discussion. It's better than what all those flat forums you mention tend to do: closing topics with aggressive admonition to use a search feature (which is near-universally broken anyway).
> They could, but rareley do. On Reddit this requires clicking many links.
It doesn't, unless you're using the new UI, at which point it's kind of your fault anyway, because Reddit is entirely unusable as a discussion place with its new UI. You have to click to load more than 3 anything, and half the time it refreshes and loses your position anyway.
If the problem is just the sheer amount of data and not navigation, you can simply bookmark the thread and come back later. The algorithm usually does a good job getting the best comments on top after things have settled down, just read these. Plus, if you forget about the topic and don't come back to it, maybe it wasn't that interesting after all, that's a good filter.