It's an acceptable trade-off to me. But I wonder if others would also be interested in longer term discussions, and if there could be a way to have them. I thought about old forums, where old threads get bumped even if they were created years ago, and wondered if an hybrid model could be of interest to HN users.
Just a thought, I'm glad to have this site as is. Enjoy your Sunday everyone!
However, our gating factor is time and attention - we only have a fixed amount. I suspect that what makes many HN discussion valuable is that the few top discussions draw most of the focus of a large part of the community at the same time. Without that concentration of focus, you don't get those spontaneous interactions where someone makes a comment about a decision made in a 30 year old piece of software powering half the Internet, and the guy that wrote said software responds with the rational for why he made that decision at the time.
Long lived threads and resurrected discussions from the past diffuse that time and attention. While I like the idea of longer term discussions (a lot!), spreading the beam of focus gets us less 'power on target' for the topics of the day.
A lot of projects get posted and there's a reply that mentions the 2 or 3 other times in the past it was posted, with links to the comments. I thinks that's one way to deal with longer term topics and updates--allow reposts over time and new discussion, with links to the past discussions if folks are curious to dig deeper into the history.
IMHO not much really needs to change, I think how reposts and such are handled right now is great.
Think about those millions of people, interacting for billions of hours online over almost two decades now. What is there to show for it at the end of the day? If even 1% of that exchange its somehow "valuable", it means there is a lost opportunity. And it will be kept thrown away in the future.
Wouldn't it be useful if we somehow could use technology to persist the "better" bits across this ever growing digital ocean? Something like wikipedia but autogenerated from diverse sources, with no claim to "truth", but rather a concise, searchable repository of whatever people are interested and are discussing online.
Now at the risk of getting downvoted, my experience is that Reddit is awful: spiteful posts, tons of short sentence replies that litter the eyes, circular meme-like self-referential wink-wink behavior (i.e. repeating the same joke with slight modification).
There are some thoughtful subs (star something codex) but overall it has a culture of people who write well but their reasoning stinks. They have strong opinions and make assertions speculatively. I can't read any real estate investment, market or nerdy thread on Reddit without encountering armchair bullshitisms from people who happen to write well but don't research anything, and respond with hostility when questioned.
So maybe this is my way of saying, don't change HN too much.
Shallow, reflexive engagement (especially from a "large" number of participants) has no actual value except to platforms needing eyeballs for ad revenue, which isn't the HN model. So.
Brought to you by the letters I, M, and O.
I used "slrn" to read news and it was great. It would only show me new posts. I could kill a thread to filter it so I didn't have to see it. I could just hit one button to go to the next post on the thread or to the next thread.
In many ways it was superior to modern web based forums.
(For example in a new proposed rocket technology, after 100 or 200 comments there will be a huge discussion about how to use it to get FTL travel like in a popular sci-fi movie.)
In lieu of a built-in solution I've been using https://www.hnreplies.com/ which seems to work well.
I did a Show HN back in April and have been quietly continuing to develop the site. Open to ideas/suggestions as well.
Is there a middle ground? I don't know. Edit : Daily/Weeky email (=slower) updates ?
As for old threads getting bumped, I think this is here equivalent to reposting the link (or re-asking). The old thread is almost "dead", but this is how it works "in real life discussions". So maybe that is not so bad. This is (as a call it) a (fast) stream model, where there is no (direct) accumulation : it is more twitter than Stack Overflow, more blogs than Wikis.
All and all, reliving prior content and/or having a complex multifunctional interface would take away from the centrality and freshness of the home page experience.
I have wondered if dang would be open to allowing community to contribute to various approved feature additions; for example, better search, night mode, duplicate detection, etc.
To encourage longer term discussion without changing the existing design, one idea is to add longer term notification to keep the engagement on a thread discussion, e.g. a daily summary of the responses to my comments coming as a notification. One is not constantly bombarded with instant notifications but still have a chance to catch up on the responses.
I remember Usenet newsgroups were sync’ed daily via uucp, which forced a delayed notification of new responses.
A long-form HN would involve another audience than the dopamine-high people.
I think you might respond that improving the site's discussion features would improve the quality of the discussion. But I am skeptical that would happen: it feels like an engineering approach rather than a community approach. The best discussion sites on the internet have worked consistently because of the crowd they attract, and the pressure moderators apply, not because they had a certain feature set.
Every year or two the same thing needs opened for conversation - in some cases.
You can always reference the old post.
Bots already pretty much dominate platforms like Twitter and Reddit, although I would point out the structured nature of Reddit makes it an ideal public resource AI training dataset. Saying that, I like your idea because these new LLM's like ChatGPT v3 need to be able to debate which is something that has largely been shutdown online including here, under a variety of guises. A debatable AI some would argue is the next logical step in some area's of AI evolution.
Past HN threads on evergreen topics do have a very long tail. I'm still getting inquiries in my area of expertise* about items I posted on HN 5 years ago. I really appreciate this and my inquirers do too.
H/T to dang for curating all this. Thank you.
*US extraordinary-ability visas for engineers / tech industry.
Many people here have contact info in their profiles. If you find a good (dead) comment thread, you can contact the people involved directly. I suspect many of the people here would be willing to continue the conversation.
The other way to restart discussion is to repost (if the topic is more than a year old). That's a pretty common theme on HN. You can link the past discussion in the comment section for context.
IIRC the BBS software surfaced threads based on recent replies rather than novelty so interesting threads stayed current for longer.
It would be hard to replicate and I am not sure how one would start but I miss it :)
HN has something kind of like this with /newcomments.
It would be similar to a forum.
I can't imagine running HN. If I had to run HN and also had to deal with 1-n week long flame wars I would delete the database, backups, and repo.
The worst part of the modern internet is how often we have to come into contact with commenters and moderators. No, I don't want to click on a monkey emoji to declare your Discord's constitution legally binding.
I've tried to put my finger on what made certain G+ discussions just tick, and it's complex and subtle.
Much of it came down to having a good host for the discussion, and among the best was the site's chief architect, Yonatan Zunger. He was well-connected within G+ and at Google (obviously), but also had a diverse set of interests, tolerated a wide range of opinion and interactions, but not without limits. (Similar in many regards to HN's moderation philosophy.)
Threads were anchored by the initial post or link, and Yonatan often wrote a fairly substantial piece himself, rather than just throwing up links or a brief tweet-length bit.
Discussions were limited in how many comments could be appended, I believe the limit was 500. Presentation was unthreaded flat, with most recent comments at the end of the thread ... but that was also what was presented by default (the full thread could be expanded if desired).
And "recent" participants (I believe this was the most recent 100 comments, could have been more) would get a notification if there were new comments. This could (and did) occur even years after the initial post, and occasionally old posts would come back to life.
Participants would receive a notification, and could view and respond to recent comments within the Notifications pane itself, rather than having to separately go to the post itself. This is unlike many other systems I've seen, including Diaspora*, which is otherwise quite similar but has a very cumbersome (and slow) notifications -> posts interaction.
Key to keeping this from becoming a magnet for spam or tedious arguments was the fact that the host could moderate (and remove) low-value, distracting, or inappropriate replies. Doing so would effectively remove notifications generated by such comments. Again, Yonatan was diligent in this --- over several years I saw a very small handful of low-value responses, and usually they had been removed in the few seconds between my own seeing them and writing Yonatan to alert him of something untoward.
The other element driving this was an effective search tool. G+ went through numerous iterations, some lacking search entirely and some with a very half-assed search, but for its final few years, there was actually an effective and useful search which could turn up interesting prior discussions. (Search is another feature Diaspora* is entirely lacking.)
As regards HN, several elements that made G+ so vibrant in this regard are lacking. There are no notifications, individual members don't host threads in the same way as G+ did, S/N is worse than G+ at its best (and that was by no means the typical experience).
For HN, the best way to resurrect an old thread is probably to either link it (or a comment) directly, or to re-share the original article/URL and reference earlier discussion. That would have to be more than just the "previous discussion" comments which are typically made.
I'll often reference my own earlier comments (or occasionally posts elsewhere) in responses, and have a few bugbears I'll raise again. (Those ... risk becoming tedious, I try to be mindful of that.)
There are a few themes I've raised multiple times and developed somewhat in the course of multiple discussions, two examples that come to mind are the prospect of trans-oceanic trains (using submerged floating tunnels), and the possibility of a tax- or ISP-brokered scheme of universal content syndication as an alternative to subscriptions and ad-supported media, both of which I've developed over several years.