3 ms·

It's ambiguous what GP refers to as "next", but if it's the Eternal September part, I believe HN unfortunately already suffers a lot from it. In my subjective o

It's ambiguous what GP refers to as "next", but if it's the Eternal September part, I believe HN unfortunately already suffers a lot from it. In my subjective opinion, comment quality took a nosedive in the past 2-3 months or so. That said, I have no idea how to fix it, if it needs fixing at all.

dang 3y ago
I'm not saying you're wrong and anyway it is hard, if not impossible, to evaluate objectively—but I can tell you two things for sure. One is that people have been saying more or less exactly this about HN for at least 15 years; the other is that HN is subject to a lot of random fluctuations, and random swings tend to get interpreted by humans as long-term trends—not because they are, but because that is what humans do.
- dredmorbius 3y ago
  In addition to my sibling comment: HN also steps in to quash developing negative patterns in all sorts of ways. There's a long list of banned sites, there is the flamewar detector (though I've ... questions ... about that), dupes detection (or flagging), there are weightings and penalties given to various sites. I believe also some keyword and other patterns are looked for as well, "Reddit" being among ones dang's recently discussed. So yes, occasionally some new pattern or trend will emerge, but HN adapts to those fairly quickly.
dredmorbius 3y ago
Paralleling what dang's said here, I've been looking at 17 years of HN front page activity over the month or so, and am starting to tackle the question of topic drift and/or focus over that period. I've been sort of live-blogging the experience on the Fediverse: <https://toot.cat/@dredmorbius/tagged/HackerNewsAnalytics https://toot.cat/@dredmorbius/tagged/HackerNewsAnalytics>, as well as in some of my HN posts. My current tack involves looking at sites (as reported in parentheses at the end of each HN front-page post title) and classifying those. With slightly more than 30% of sites categorised, I can classify about 65% of all HN posts. For the full dataset (17 years), that's roughly: 1 63913 35.73% UNCLASSIFIED 2 22589 12.63% blog 3 15112 8.45% general news 4 13823 7.73% tech news 5 12851 7.18% programming 6 8622 4.82% corporate comm. 7 8459 4.73% academic / science 8 7294 4.08% n/a 9 5324 2.98% business news 10 3803 2.13% general interest 11 2151 1.20% social media 12 2074 1.16% software 13 1613 0.90% technology 14 1463 0.82% video 15 1144 0.64% general info (wiki) 16 1009 0.56% government 17 724 0.40% misc documents 18 720 0.40% law 19 702 0.39% tech discussion 20 620 0.35% science news Tons of caveats: this depends heavily on how I classify individual sites, a given site's stories might well be technical, social, or political, etc., etc. The breakdown-by-year analysis is in development, but if anything programming-specific content as increased in prevalence. Political discussion seems not to have (though it rose significantly ~2014). Cryptocurrency and blockchain-specific sites also peaked about that time (I suspect much of that discussion is now mainstream). General news has always been a huge portion of HN discussion, as have individual (and corporate) blogs. Note again that this isn't about discussion and comments, or even the titles or article contents (I'm thinking of looking at those, it's ... a challenge for me). But across nearly 200,000 front-page stories, on which nearly half of all HN discussion occurs (based on another API-based study looking at comprehensive posts), the overall trending seems at first blush to be pretty consistent and if anything improving over time. (As with all preliminary results, I'm hoping I won't have to eat my words here. Though I'm reasonably confident in most of this.) From the classifications above, the places you might find some that "suffering" would be in general news, genral interest, and social media categories. All but the first of those are single-digit percentages, and a lot of that general-news content is about technology, business, finance, and science, all of which would crowd out the sort of social and political issues which seem to generate strong feelings. The "UNCLASSIFIED" sites are a wide mix, though most are probably a mix of blogs, corporate / organisational communications, and the like. The mean posts per site is 1.739951, so gains from additional site-categorisation are pretty slim. I have captured a lot of obvious patterns via regexes and string matches, so academic/science and major (or even minor) blogging and social media sites aren't a large fraction. More recent discussion here: <https://news.ycombinator.com/item?id=36524001 https://news.ycombinator.com/item?id=36524001>
- dredmorbius 3y ago
  More on "UNCLASSIFIED": there are 36,520 of those sites. It's not practical to list all of them. But we can randomly sample. And large-sample statistics start to apply at about n=30, so let's just grab 30 of those sites at random using `sort -R | head -30`: 1 sfg.io 1 extroverteddeveloper.com 2 letmego.com 1 thestrad.com 2 bombmagazine.org 1 domlaut.com 1 bootstrap.io 1 jumpdriveair.com 2 desmos.com 1 leo32345.com 1 echopen.org 1 schd.ws 1 web3us.com 7 akkartik.name 1 bcardarella.com 1 cancerletter.com 1 platinumgames.com 1 industrytap.com 2 worldoftea.org 1 motion.ai 1 vectorly.io 2 enterprise.google.com 1 lift-heavy.com 1 davidpeter.me 1 panoye.com 3 thestrategybridge.org 2 fontsquirrel.com 1 kettunen.io 1 moogfoundation.org 2 elekslabs.com That's a few foundations, a few blogs, a corporate site (enterprise.google.com), and something about tea, all with a small number of posts (1--7). I'm looking at some slightly larger samples (60--100) here on my own system, and can actually make some comparisons across samples (to see how much variance there is) which can give some more information on tuning what I would expect to find under the "UNCLASSIFIED" sites.

dang 3y ago

I'm not saying you're wrong and anyway it is hard, if not impossible, to evaluate objectively—but I can tell you two things for sure. One is that people have been saying more or less exactly this about HN for at least 15 years; the other is that HN is subject to a lot of random fluctuations, and random swings tend to get interpreted by humans as long-term trends—not because they are, but because that is what humans do.

dredmorbius 3y ago
In addition to my sibling comment: HN also steps in to quash developing negative patterns in all sorts of ways. There's a long list of banned sites, there is the flamewar detector (though I've ... questions ... about that), dupes detection (or flagging), there are weightings and penalties given to various sites. I believe also some keyword and other patterns are looked for as well, "Reddit" being among ones dang's recently discussed. So yes, occasionally some new pattern or trend will emerge, but HN adapts to those fairly quickly.

dredmorbius 3y ago

Paralleling what dang's said here, I've been looking at 17 years of HN front page activity over the month or so, and am starting to tackle the question of topic drift and/or focus over that period. I've been sort of live-blogging the experience on the Fediverse: <https://toot.cat/@dredmorbius/tagged/HackerNewsAnalytics https://toot.cat/@dredmorbius/tagged/HackerNewsAnalytics>, as well as in some of my HN posts. My current tack involves looking at sites (as reported in parentheses at the end of each HN front-page post title) and classifying those. With slightly more than 30% of sites categorised, I can classify about 65% of all HN posts. For the full dataset (17 years), that's roughly: 1 63913 35.73% UNCLASSIFIED 2 22589 12.63% blog 3 15112 8.45% general news 4 13823 7.73% tech news 5 12851 7.18% programming 6 8622 4.82% corporate comm. 7 8459 4.73% academic / science 8 7294 4.08% n/a 9 5324 2.98% business news 10 3803 2.13% general interest 11 2151 1.20% social media 12 2074 1.16% software 13 1613 0.90% technology 14 1463 0.82% video 15 1144 0.64% general info (wiki) 16 1009 0.56% government 17 724 0.40% misc documents 18 720 0.40% law 19 702 0.39% tech discussion 20 620 0.35% science news Tons of caveats: this depends heavily on how I classify individual sites, a given site's stories might well be technical, social, or political, etc., etc. The breakdown-by-year analysis is in development, but if anything programming-specific content as increased in prevalence. Political discussion seems not to have (though it rose significantly ~2014). Cryptocurrency and blockchain-specific sites also peaked about that time (I suspect much of that discussion is now mainstream). General news has always been a huge portion of HN discussion, as have individual (and corporate) blogs. Note again that this isn't about discussion and comments, or even the titles or article contents (I'm thinking of looking at those, it's ... a challenge for me). But across nearly 200,000 front-page stories, on which nearly half of all HN discussion occurs (based on another API-based study looking at comprehensive posts), the overall trending seems at first blush to be pretty consistent and if anything improving over time. (As with all preliminary results, I'm hoping I won't have to eat my words here. Though I'm reasonably confident in most of this.) From the classifications above, the places you might find some that "suffering" would be in general news, genral interest, and social media categories. All but the first of those are single-digit percentages, and a lot of that general-news content is about technology, business, finance, and science, all of which would crowd out the sort of social and political issues which seem to generate strong feelings. The "UNCLASSIFIED" sites are a wide mix, though most are probably a mix of blogs, corporate / organisational communications, and the like. The mean posts per site is 1.739951, so gains from additional site-categorisation are pretty slim. I have captured a lot of obvious patterns via regexes and string matches, so academic/science and major (or even minor) blogging and social media sites aren't a large fraction. More recent discussion here: <https://news.ycombinator.com/item?id=36524001 https://news.ycombinator.com/item?id=36524001>

dredmorbius 3y ago
More on "UNCLASSIFIED": there are 36,520 of those sites. It's not practical to list all of them. But we can randomly sample. And large-sample statistics start to apply at about n=30, so let's just grab 30 of those sites at random using `sort -R | head -30`: 1 sfg.io 1 extroverteddeveloper.com 2 letmego.com 1 thestrad.com 2 bombmagazine.org 1 domlaut.com 1 bootstrap.io 1 jumpdriveair.com 2 desmos.com 1 leo32345.com 1 echopen.org 1 schd.ws 1 web3us.com 7 akkartik.name 1 bcardarella.com 1 cancerletter.com 1 platinumgames.com 1 industrytap.com 2 worldoftea.org 1 motion.ai 1 vectorly.io 2 enterprise.google.com 1 lift-heavy.com 1 davidpeter.me 1 panoye.com 3 thestrategybridge.org 2 fontsquirrel.com 1 kettunen.io 1 moogfoundation.org 2 elekslabs.com That's a few foundations, a few blogs, a corporate site (enterprise.google.com), and something about tea, all with a small number of posts (1--7). I'm looking at some slightly larger samples (60--100) here on my own system, and can actually make some comparisons across samples (to see how much variance there is) which can give some more information on tuning what I would expect to find under the "UNCLASSIFIED" sites.