Hacker News new | past | comments | ask | show | jobs | submit login
Hacker News Classics (2018) (jsomers.net)
165 points by katzeilla on May 10, 2021 | hide | past | favorite | 42 comments



I've long had this notion that there are two kinds of content on a site like HN:

- temporal, like news, announcements, analysis and discussion on current events. It's usually irrelevant after a while, it gets stale fast.

- atemporal, like essays, history, theory, articles, knowledge in general that while it may get superseded, expanded or invalidated over time, it's interesting content that you learn from, from a current or historical perspective.

While the first kind tends to get more attention due to the clickbaity nature, I much prefer the second kind and wonder if a site focused solely on that would be more interesting.


I know where you're coming from, but I think the shallower current events type discussions bring a lot of value in holding a mirror up to the deeper, more theoretical and timeless pieces.

It's about keeping them honest. You know the latter have value if you see their themes and lessons play out frequently in the former.

A decent test of the most useful theory is how often it's relevant and referenced in shallow quick fire current events discussions.


A similar concept is the book serie 'A Map that Reflects the Territory' : https://www.lesswrong.com/books.

It's a selection of essays that were written for the LessWrong community, and the community voted for the best essays 2 years after publication, ensuring that only the 'atemporal' ones got into it.


You can generalise this to much document-style information (in the Paul Otlet sense).

Nothing has the lack of persistence that news does. (Though a surprising amount of "news", for reasons closely related to this rapid-staling and unpredictable nature of it, is actually repackaged evergreen or procedural content.)

At the same time, there's little that's as timeless as highly-insightful observations or commentary. These can live on for years, decades, centuries, or millennia.

(I suspect that copyright attacks both ends of this equation, by creating moats around current material to keep it in flow, and by denying markets to older material, which would otherwise compete with the new. This has limits imposed by works which were in fact subject to copyright in the first place, which is most works published since January 1, 1926, in the US.)

I'm finding classic works, philosophy, and much academic publishing to be far more interesting than news, of late.


I feel much the same way but I'd like to add some support for having the former type of content on HN.

You do get a large response of unqualified opinions, anecdotes and repetitive arguments/flamewars but, amongst the 'noise', you can roughly cluster voices, average out within clusters and get an idea of the range of (mostly) intellectual opinions held.

This 'fuzzy' information won't be neatly laid out as with the atemporal articles but to me it acts almost like a primary resource surveying what the HN community (whatever that represents in reality) believes at a point in time.


Yes. I think this also aligns with the origins of HN as "startup news" (news is temporal), which was expanded to "gratifies one's intellectual curiosity" (timeless) https://news.ycombinator.com/newsguidelines.html

Suggest: two sites


Agree, and while the former might best be represented on a dated "timeline", esp as an ongoing thing of related events; The latter is maybe best as a "slow journalist"/"community wiki" style of aggregating useful comments.


MetaFilter might be for you then


Not to be confused with Hacker News' /classic: https://news.ycombinator.com/classic

/classic uses the same ranking system as the regular front page but only counts votes from early users, defined as accounts that were created before Feb 13, 2008.

I'm guessing this is an attempt at getting around the Eternal September phenomenon, or at least an experiment around this idea.

Source: https://news.ycombinator.com/item?id=24401292


And I was thinking "Hacker News Classics" would refer to well-known posts/comments on HN, like the comment of the guy completely dismissing Dropbox.


This is very inline with the Lindy Effect - things that have been around for a while can be expected to be relevant longer than things that were just created yesterday. Our information diet is only as useful as is relevant in future, and tends toward too fleeting.

So I think people should make Lindy Libraries - reading resources that have stood the test of time. Here are a few I have been collecting: https://codingcareer.circle.so/c/lindy-library


> Our information diet is only as useful as is relevant in future

I think there are situations where ephemerous knowledge is useful. For example, knowing about a security vulnerability: its profitability is much higher before the issue gets fixed, which can happen anytime (be it two hours or two years).

Another example: imagine you know insert-influential-billonaire is going to tweet about insert-trendy-currency, this information is completely irrelevant for the future, but you can certainly use it to make money.

It's worth mentioning these two examples, however, can provide longer-lasting knowledge as a by-product too.


I know everyone is wondering when Hacker News Classics (2018) will appear on Hacker News Classics (2018). From the underlying code it looks like this depends on: (a) fetch.rb to be run again to regenerate stories.json; (b) the definition of “classic” to be expanded beyond 1900–2010; (c) the Algolia HN API to update which it hasn’t yet. [1] [2]

[1]: https://github.com/jsomers/hacker-classics

[2]: https://hn.algolia.com/api/v1/search?tags=story&query='2018'...


Unfortunately some dead links, e.g. https://news.ycombinator.com/item?id=6512288

Would be cool if HN had its own web archive after 100 comments or so.


Or a bot that browses discussions and if any recent ones have > 100 comments automatically sends the link to archive.org to be archived and then add a comment with the link of the archived copy


Are bots allowed here? I've never seen the kind of utility bot that is common on Reddit.


Bots are heavily discouraged here. (Almost totally forbidden.)

Perhaps if the bot sends the page to the archive immediately but instead of posting the comment immediately waits a few days (until the last day comments are allowed), it will not so controversial because it will add no noise to the conversation.

It's better to send an email to the mods with the idea before implementing it, because the bot may get banned.


Found a replacement for the (broken) video link in "WATCH THE VERY FIRST FILM VERSION OF ‘ALICE IN WONDERLAND’ FROM 1903":

https://youtu.be/zeIXfdogJbA


collecting stories that have (year) on them once may have been a nice way to surface archive essays and interesting pieces. These kinds of posts are submitted daily now and it's getting annoying/ridiculous when things are submitted less than a year apart, sometimes even monthly!

Old articles that have been discussed heavily, improved upon, circumstances changed, technology evolution especially, are mostly lame to revisit on a site that still should weigh heavily towards current events/news developments.

I'm just getting tired of sifting thru resurfaced old posts from 2016 or whatever


I enjoy this idea. Always nice to look at the "classics."

I wonder if maybe the Y logo top left should be removed as it might mislead users into thinking it's a legitimate extension of HN?


It would be cool to have a version sorted by (parenthesized) date.

Oh, and "(1900..2010).each do |year|" is not good: https://news.ycombinator.com/item?id=16444460


That's fantastic.

"Linux is obsolete (1992)": https://news.ycombinator.com/item?id=8942175


Linus says:

"If the GNU kernel had been ready last spring, I'd not have bothered to even start my project: the fact is that it wasn't and still isn't. Linux wins heavily on points of being available now."

Is he referring to GNU Hurd that is still unfinished?


Correct - the general assumption at the time is that GNU Hurd was about to bust out and become the main free software kernel. Remember that computers that COULD run practical Unix for home use had just become relatively available - and Linus wanted to play with his new 386 and the fun instructions provided therein.


> Linus "my first, and hopefully last flamefest" Torvalds


Some more discussion from a year ago: https://news.ycombinator.com/item?id=22004066


Yup - plus an earlier:

Hacker News Classics (2018) - https://news.ycombinator.com/item?id=22004066 - Jan 2020 (42 comments)

Show HN: Hacker News Classics - https://news.ycombinator.com/item?id=16442888 - Feb 2018 (107 comments)


A somewhat related shameless plug: Timeless Hacker News https://thn.rakhim.org/


Half-OT:

HN could optimize the layout for mobile a bit. I'm mostly reading on my smartphone and the links are all sooo small, I can barely hit them with my thumb.


I use http://hn.premii.com/ on mobile. No support for commenting but for just browsing is very nice IMO.


You could use Materialistic or any other apps that make it easier to view HN.


Unfortunately HN locks the comments after a while.

Mods: we need a way to fix deadlinks

Example: https://news.ycombinator.com/item?id=1140283 redirects to wrong place.

Correct new link is https://lockhaven.edu/~dsimanek/hollow/tamarack.htm


Dang has said this is done as a way to preserve history. I think that’s more important than some of the links being outdated.


That's generally true, but if it's a historical source that's still available elsewhere, then the particular link doesn't have so much historical value.

Occasionally I'll post archive.org links or similar alternatives as pinned top comments, so readers looking at the old thread can have access to the article. In the past we've changed the top link to archive.org also, but I'd rather not encourage that.


We updated that one — thanks!


It is not on https


The submission link needs HTTPS. But the website itself supports HTTPS.


> the web is the greatest library in the history of the world

I think this is web-boosterism. Consider another candidate for greatest library in history of the world: the British Library. 200 million items, most of them available to read at 45 minutes notice, centrally located in one of the worlds most important cities.

I'm bet that 99% of what is on this page is in the BL. Is 99% of what is in the BL on the web?

EDIT: For context, here is some of the stuff in the BL: https://en.wikipedia.org/wiki/British_Library#Highlights_of_...


How do I access the British Library from India? Or perhaps from Australia? New Zealand? What about Canada?

The Internet offers a unique value proposition the British Library does not: it's globally available, for free, and in the case of Wikipedia might even come in your native tongue. It also offers video content, live or otherwise, on any Internet connected device you happen to own.

The British Library doesn't compare, but it sure is an exceptional resource.

a lot of the time


> How do I access the British Library from India?

You clearly can't (though India no special shortage of excellent libraries). But how important is access? Is the chip shop down the road from you the Worlds Greatest Restaurant just because it's the easiest to get to? Or is that place in Paris with 3 Michelin stars better?

The web-as-a-library relies on people creating good stuff and paying for it to be on the web for free. It just doesn't happen that often and Wikipedia is, I'm afraid, not the last word on human knowledge.


Well, the web has some obvious advantages over mostly-paper libraries.

I'm sure a certain library in Alexandria would have appreciated off-site backups :)


> centrally located

That's not necessarily a good thing. Think of a fire.




Join us for AI Startup School this June 16-17 in San Francisco!

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

Search: