I'm actually really interested in this addition. I'm just going back to 'big' days, such as financial crash. When I started HS when I finished HS, when I started College. A few articles are interesting, it's definitely better than bumbling around Algolia.
Edit: One great link is this one
> Google Making Extraordinary Counteroffers To Stop Flow Of Employees To Facebook [1]
I like it. It seems to surface more unusually good stories -- but is that a problem?
Won't the list end up dominated by a handful of winner-take-all stories, lingering throughout the entire day at the top of the list, rather than an ongoing series of slightly-less-popular stories popping in and out?
If there's always going to be better stories with the "sorted by time spent" view, does that create a danger that everyone will just devote their attention to those few high-grade stories, and ignore all of the rest?
It's just a list of past front pages. Once each day is over, its front page never changes. Does that answer your question?
(Well, it's not exactly a list of past front pages, because what appears on /front was never the actual front page at any point. Since that's always changing, we take a kind of average.)
Yeah, I was confused about how this works. But if the link appears at the bottom of the page, then this new page will be more of an afterthought (after reading the current front page and upvoting its stories at that moment in time).
Edit: Oh, also: "more" fits with the "More" links you see at the bottom of the front page and other places, clicking which takes you back to previous posts.
Edit 2: Ok, you guys have convinced me you don't like "more", so let's go with "past". That word is overloaded with 'past' links that point to HN searches, but perhaps those meanings are compatible.
It's good for personal bookkeeping. Whether you've already read a story, the comments or just have no interest - it's nice to get off your page so you don't have the unnecessary clutter and/or distraction.
BTW: If there was something to reduce accidental hits on hide and flag, that would be great.
On mobile touch screens, being right under the main links they can get hit a surprising number of times and then you've got to go fumble around to undo it.
"Afore" would be nice, but it's too esoteric.
How about "Ahead" ?
Synonym of previously, as well as goes with the idea, like "Go ahead and read another story"
Seems like an archive feature? Since you can go back to see the contents of front pages from previous days. But "Front Page Archive" doesn't necessary condense well into single word unless you're going to call it /archive.
Agree. More is too often used to display 'more options', which is what I thought this was going to be. I tailgated on your Snapshot and "Time Capsule" suggestions (which I liked) to try to find other one word alternatives. Came up with the following:
cache, record, timecard, log, chronicle, registry, diary, chronology, timeline, calendar, register, log and archive.
I'm partial to snapshot, archive and timeline - as one word substitutes.
more could be confused with the more link at the bottom and it doesn't really tell the whole story.
i discovered front by accident when the above announcement came up in a search, and i have been using it ever since to keep up with the top hits of the day. it actually helps me to read less and limit my time on hackernews.
Personally, I would call it "time" as a dual-meaning single word that references both the ability to go to a specific day and the fact that it is computed based on time spent (and maybe also "go here if you find yourself with some extra time to sink in").
it may look odd, but hackernews readers are inquisitive, and will check it out. on the other hand i would guess that most will ignore more thinking it's the same as the more link at the bottom.
actually the problem with front is not on the front page but on all others where it will make people think that it goes to the front page...
Only now I realized that I use https://hckrnews.com for so many years that I have come to think of it as the actual HN front page. It's hard to imagine something better.
Yes! I love https://hckrnews.com, and the ability to come back later in the day and catch up on anything that was on the front page, and not worry about missing out.
I also echo the sentiment of being used to that site as the HN homepage.
It can also show the stories from arbitrary periods including single days like in the new "more" feature being discussed but it orders them by points rather than how long they were on the (real) front page.
Lately, I have a feeling that hckrnews surfaces the wrong type of posts - stuff that quickly reaches 3 points. But it's often flamebait, clickbait or some useless drama. I often go back to the home page to find posts that don't appear on hckrnews but are quite interesting in a niche or need to be fully digested to appreciate.
That said I also love hckrnews for checking what's new after a longer time
I feel that the top navigation "more" link could be renamed into "popular by date" or "yesterday's popular".
Those labels appear more accurate to me and better convey the functionality of the current top "more" link.
Moreover having two links with the same name on the same page which behave differently seems misleading.
In fact, I was wondering why someone would want to change the 'More' link at the bottom of the page. Took me a moment to realize there is another one in the header now.
I love that HN doesn't change for change's sake. One problem I have with tech orgs is that inevitably their designers will find work to do, necessary or not.
So when things do change, I like that it's very methodical and usually quite minor.
Methodical or just overly cautious? It took ages to get the `[-]`/collapse tree link, and the other IMO obvious QoL improvement -- a 'parent' link to a) get to the root node of a discussion tree that no longer interests the reader and then to b) use the mentioned '[-]' to collapse it all -- might take even longer.
I like the "Time Machine" feature of going back 1 (to N) year(s) in the past, it brings a lot of insight as to how technologies and trends evolved (for those of us who have only recently joined Hacker News).
This is one area when the functionally can really be improved over what reddit implements. Having a moving window of time to sort within is an amazingly useful feature for sorting out trends or lost articles. The expanded granularity would be very much appreciated.
I don't understand what I'm looking at, other than the option to view the page as it appeared a day, month, or year back... or a day in the future somehow?
I'm curious how the algorithm works, as it doesn't seem to work as I'd expected.
My first thought was to go back and look at the day I launched my startup with at Show HN [1], which got over 700 points and was at #1 from roughly noon to 2 AM the next day (PT). Sure enough, this post shows up in the #2 slot on the day I posted it (Sept 5, 2013) [2], but it doesn't show up at all on the following day [3].
This is strange because it was at #1 at the beginning of this day (assuming that's midnight PT) and was on the front page for much of that day. Even when you go hundreds of entries down into the Sept 6 page, it doesn't show up.
Any idea how the algorithm works that it's generating these results?
We only started logging the front page on 2014-11-11. For days before that, what you're seeing is the stories which were submitted on that day, ordered by vote score. (Not necessarily publicly visible points, though, because some votes get dropped for various reasons.) Since your story was submitted on Sept 5, 2013, that's the day it shows up.
That’s certainly a possibility, and it’s quite possible that the current algorithm either has more gravity or is not as favorable to Show HN posts. Either change would explain why this particular post would be treated as if it hadn’t been on the front page at all on the day after posting.
If this is the case, it should be clarified that this feature uses the current algorithm, which may not reflect how things ranked on a given day.
Any plans to update the official Hacker News API? I’d love more features, such as ability to reply and upvote, but more importantly, the API currently requires using a really old version of Firebase SDK. The rest API is an alternative, but it doesn’t work well due to the design of the API needing a whole network call per post/comment.
I’d say allow at least 3 months overlap between releasing the JSON API and killing the Firebase API. 6+ months would be ideal. Lots of apps need updating!
I'm curious what's the reasoning behind adding this link? To me it makes HN much easier to catch up. I see this as the "slow media" movement: https://en.wikipedia.org/wiki/Slow_media
I like to explore it more visually. https://hnews.xyz/
Is that something you might consider adding in the future?
( I’m the author of https://hnews.xyz )
I wish there was a hn viewer integrated with Outline. something like, for each new article on front, check whether it can be outlined and, if it can, just display it without redirecting to the original site. That would protect against slashdotted sites, paywalls, annoying gdpr consent pop ups, sorry but we don't like europe pages, all kinds of tracking and js heavy websites (especially if you're on i.e. Lynx) etc. This could be done either by redirecting to the Outline website or using its API (see the network traffic for details on how it works). Also a HN that would be essentially a list of check boxes ordered by time and an "add selected to Pocket" button would be nice too.
Hypothesis is the company who makes the open highlighter / web annotation software on Outline. [I believe they also run Outline but I can't find this made apparent anywhere. It could be that Outline just uses the Hypothesis annotator client.]
If you register for a free account on Hypothesis (sometimes stylized Hypothes.is), you can leave private or public highlights and annotations while reading on Outline, and if you install their browser extension on any webpage.
Genius (YC S11) (formerly Rap Genius) started down the path of creating a global web annotator but pivoted into a media company to monetize. Hypothesis is basically what Genius could have become if they continued with the vision of annotating the entire web and fully built out the product. The company itself is a non-profit and their code is open source.
If I were them, I wouldn't publish the fact that they're running outline very widely considering the dirty tricks that service employs, i.e. to show wsj articles to non-subscribers etc. I like the fact that someone finally has the guts, but if I were them, I wouldn't reveal myself that easily.
It would be cool to see links from previous days/months/years point to an archived version of the link (if possible). If you go back a substantial amount of time, you end up hitting a lot of broken links.
Could be a quick Greasemonkey/Tampermonkey script hack to grab the date from the URL and links to the nearest Wayback Machine version using https://archive.org/help/wayback_api.php
I think the UI is kinda awkward to use. If I want to input a specific date I have to manipulate the URL? Gets really annoying with mobile. Also it would be nice to somehow see the past dates somehow visualized. Maybe how much total activity there was per day, or what types of topics were discussed. Now I am just going blind without any sense of direction what lies ahead. It's very refreshing feature and all but I will probably only use it for past day's submissions since going farther fishing for interesting submissions is too laborious.
I wrote a browser extension a while ago that adds more of these lesser known HN pages to the nav. I like having a way to get to all of them individually from the top of every page.
This is cool, thank you! Would “days” be a good name as it allows you to browse a set of days?
It doesn’t seem like posts you’ve hidden from the front page are reflected here. I guess they don’t really need to be since this is more archival in nature.
This is great. I always wished I could see front pages from the past.
The only downside is that getting bored of the current front page is sometimes the only reason I get off HN when I'm procrastinating and this makes it easier to prolong that. :)
see hnrss.org. I use it along with an iOS shortcut that retrieves the articles for me, concatenates to a big txt file, uploads to dropbox so that I can read later on whatever I want and automatically starts reading with the build-in tts.
In the first few months, there were quite a few days where nothing was posted (see the difference between https://news.ycombinator.com/item?id=61 and 62). If /front works for a day, but shows no stories, that means HN existed but there were no stories that day.
The navigation links work until they point outside the ultimate beginning or end of the data, and then stop appearing. That was a bit tricky to get right as I recall.
I hear you, but proper timezone support is one of those things we've shied away from over the years. It's hard to get right and our dev resources are too limited. So mostly we just work with UTC. I'm sorry—I know it can create a bit of an outsider feeling, which is definitely not intended.
Yeaaaaah but it was on topic, in a chain of threads making fun how unsubstantive people’s skepticism was back in 2010
Your team tells me that upvotes/downvotes are noisy, but once you have downvotes it is unsubstantive and the account is rate limited
Pick
Your team also never explains how you mods interpret a guideline or substantiation, just unilaterally say it violates and pray for classical conditioning as if you want community standards to be decided by the community
If you are judging this primarily by flags and downvotes, you arent noticing that a lot of the collective conscious is just incorrect. It isnt as open to ideas as you might think
I'm happy to explain how we interpret specific cases, but not as part of a litigious process in which explanation only leads to more objections. If you want a serious answer, you'll need to do more to establish good faith.
I largely browse using the front page. I am not looking to keep up with every article ever, I just look for a few interesting things to read, then close it and move on with my day.
I also find the design and web layout to be very pleasing.
My RSS reading habits never recovered from GR's demise... but maybe one of these days.
"I am not looking to keep up with every article ever"
HN's RSS feed does not post every article ever, but just those that make it to the front page.
It's an alternate way of accessing the same articles, but I find it much more powerful and useful than using a web browser for the following reasons:
1 - I have a local archive of the titles of and links to each article that does not depend on HN. So if HN ever goes away, I can still access them all.
2 - I don't need to click through and wait for multiple web pages to load to browse through my archives.
3 - Newsboat, the terminal-based RSS reader I use is far faster and more responsive than any web browser, on my old, slow laptop.
4 - Newsboat is also way more powerful and extensible in ways that a plain web page is not. I can easily do things like tagging and filtering, I can delete articles that I don't want to see while saving the ones I do without actually having to open up the destination web page, and a lot more.
5 - I can locally search through articles without telling any external search engine company what I'm searching for. This is a small but important way to practice defense-in-depth in the realm of privacy.
At some point I'll probably try switching to reading RSS feeds from within Emacs, which should be even more flexible and powerful.
Edit: One great link is this one
> Google Making Extraordinary Counteroffers To Stop Flow Of Employees To Facebook [1]
[1](https://news.ycombinator.com/item?id=1653702)
>That's got to be a joke.
Love it!
Worth remembering these little snippets of history next time someone pooh-poohs your idea.
Won't the list end up dominated by a handful of winner-take-all stories, lingering throughout the entire day at the top of the list, rather than an ongoing series of slightly-less-popular stories popping in and out?
If there's always going to be better stories with the "sorted by time spent" view, does that create a danger that everyone will just devote their attention to those few high-grade stories, and ignore all of the rest?
(Well, it's not exactly a list of past front pages, because what appears on /front was never the actual front page at any point. Since that's always changing, we take a kind of average.)
https://news.ycombinator.com/item?id=12073675
https://news.ycombinator.com/item?id=18512028
Edit: Oh, also: "more" fits with the "More" links you see at the bottom of the front page and other places, clicking which takes you back to previous posts.
Edit 2: Ok, you guys have convinced me you don't like "more", so let's go with "past". That word is overloaded with 'past' links that point to HN searches, but perhaps those meanings are compatible.
I rejected it originally because there are already 'past' links on thread pages that point to HN Search, but perhaps those meanings are compatible.
So I will suggest: Date
Alternately, I will suggest Top since it defaults to Stories from (today's date), ordered by how long they were on the front page.
Hopefully if enough people toss out suggestions, one will turn out to be a real improvement over more.
Maybe Explore?
On mobile touch screens, being right under the main links they can get hit a surprising number of times and then you've got to go fumble around to undo it.
Maybe a "comfortable" setting a la Gmail?
Just kidding (mostly).
I think "more" is a confusing choice because you already have a "More" link at the bottom of the page that goes to the next page.
- By Date
- Date
- Rewind
- History (confusing?)
- Archive
- Old (new | old)
/runs
More more doesn't seem so bad? It's between more and more more more.
If anyone can figure out how to make it shorter, let me know. It feels a bit wordy now.
"...ordered by {time spent there}->(how long it was on the front page)"
Thinking.
Really cool concept in my opinion; no matter what the name is.
cache, record, timecard, log, chronicle, registry, diary, chronology, timeline, calendar, register, log and archive.
I'm partial to snapshot, archive and timeline - as one word substitutes.
more could be confused with the more link at the bottom and it doesn't really tell the whole story.
i discovered front by accident when the above announcement came up in a search, and i have been using it ever since to keep up with the top hits of the day. it actually helps me to read less and limit my time on hackernews.
actually the problem with front is not on the front page but on all others where it will make people think that it goes to the front page...
I also echo the sentiment of being used to that site as the HN homepage.
https://hn.algolia.com/?query=&sort=byPopularity&prefix&page...
It can also show the stories from arbitrary periods including single days like in the new "more" feature being discussed but it orders them by points rather than how long they were on the (real) front page.
That said I also love hckrnews for checking what's new after a longer time
I believed it to be more popular but it seems no one has mentioned that site so far...
So when things do change, I like that it's very methodical and usually quite minor.
Out of curiosity : Under a particular day's story, why do we see stories from the next day too ?
Eg : https://news.ycombinator.com/front?day=2018-02-21 This shows some post from 02-22 too.
Edit: it's because the /front pages were dividing up days differently than the story timestamps, which use UTC. I've made everything use UTC now.
https://news.ycombinator.com/
and
https://news.ycombinator.com/front
I didn't know the latter existed until right now.
I think the message on top takes too much space and would prefer something like
<y <d <m (2019-02-22) d> m> y>
or
<<< << < (2019-02-22) > >> >>>
or
< < < (2019-02-22) > > >
But that's just me.
My first thought was to go back and look at the day I launched my startup with at Show HN [1], which got over 700 points and was at #1 from roughly noon to 2 AM the next day (PT). Sure enough, this post shows up in the #2 slot on the day I posted it (Sept 5, 2013) [2], but it doesn't show up at all on the following day [3].
This is strange because it was at #1 at the beginning of this day (assuming that's midnight PT) and was on the front page for much of that day. Even when you go hundreds of entries down into the Sept 6 page, it doesn't show up.
Any idea how the algorithm works that it's generating these results?
1: https://news.ycombinator.com/item?id=6335784
2: https://news.ycombinator.com/front?day=2013-09-05
3: https://news.ycombinator.com/front?day=2013-09-06
If you look at https://news.ycombinator.com/front?day=2014-11-11 vs. https://news.ycombinator.com/front?day=2014-11-10, you'll notice that the text description changes at the top.
It would be interesting to see how everything is being cached for a particular day, though.
If this is the case, it should be clarified that this feature uses the current algorithm, which may not reflect how things ranked on a given day.
I have a few dozen HN apps (yours included) saved in this Pinboard collection for those looking for more alternative clients.
https://pinboard.in/u:tedmiston/t:hacker-news/
I am a heavy Instapaper user but really liking Outline / Hypothesis lately.
If you register for a free account on Hypothesis (sometimes stylized Hypothes.is), you can leave private or public highlights and annotations while reading on Outline, and if you install their browser extension on any webpage.
Genius (YC S11) (formerly Rap Genius) started down the path of creating a global web annotator but pivoted into a media company to monetize. Hypothesis is basically what Genius could have become if they continued with the vision of annotating the entire web and fully built out the product. The company itself is a non-profit and their code is open source.
https://web.hypothes.is/
https://github.com/hypothesis
https://web.hypothes.is/blog/a-letter-to-marc-andreessen-and...
It's MUCH to easy to miss interesting stories if you don't check the site at least twice a day.
Just my 2 cents!
https://github.com/skibz/hnlists
I like the function, but some how "by date" feels more "more" than "more" does.
It doesn’t seem like posts you’ve hidden from the front page are reflected here. I guess they don’t really need to be since this is more archival in nature.
The only downside is that getting bored of the current front page is sometimes the only reason I get off HN when I'm procrastinating and this makes it easier to prolong that. :)
https://news.ycombinator.com/front?day=2007-02-18
https://news.ycombinator.com/front?day=2006-10-09
https://news.ycombinator.com/announcingnews.html
Have you found the new placement to be affecting the site/database load at all?
I searched Google for "site:news.ycombinator.com features" and limited the search to past week.
[1]https://news.ycombinator.com/lists
In the first few months, there were quite a few days where nothing was posted (see the difference between https://news.ycombinator.com/item?id=61 and 62). If /front works for a day, but shows no stories, that means HN existed but there were no stories that day.
The navigation links work until they point outside the ultimate beginning or end of the data, and then stop appearing. That was a bit tricky to get right as I recall.
AHAHHA this is the same as people calling BS on today’s 300k+ annual compensation packages
Its like aw you wanted a source? Congrats you missed the boat, have fun citing things in academia
So good
We detached this comment from https://news.ycombinator.com/item?id=19230951 and marked it off-topic.
Your team tells me that upvotes/downvotes are noisy, but once you have downvotes it is unsubstantive and the account is rate limited
Pick
Your team also never explains how you mods interpret a guideline or substantiation, just unilaterally say it violates and pray for classical conditioning as if you want community standards to be decided by the community
If you are judging this primarily by flags and downvotes, you arent noticing that a lot of the collective conscious is just incorrect. It isnt as open to ideas as you might think
Such as Karma over 1,000? What do you mean
I think I’m doing pretty good, but what do you have in mind
If you don't want to be banned, you're welcome to email hn@ycombinator.com and give us reason to believe that you'll follow the rules in the future.
https://news.ycombinator.com/newsguidelines.html
https://news.ycombinator.com/rss
I also find the design and web layout to be very pleasing.
My RSS reading habits never recovered from GR's demise... but maybe one of these days.
HN's RSS feed does not post every article ever, but just those that make it to the front page.
It's an alternate way of accessing the same articles, but I find it much more powerful and useful than using a web browser for the following reasons:
1 - I have a local archive of the titles of and links to each article that does not depend on HN. So if HN ever goes away, I can still access them all.
2 - I don't need to click through and wait for multiple web pages to load to browse through my archives.
3 - Newsboat, the terminal-based RSS reader I use is far faster and more responsive than any web browser, on my old, slow laptop.
4 - Newsboat is also way more powerful and extensible in ways that a plain web page is not. I can easily do things like tagging and filtering, I can delete articles that I don't want to see while saving the ones I do without actually having to open up the destination web page, and a lot more.
5 - I can locally search through articles without telling any external search engine company what I'm searching for. This is a small but important way to practice defense-in-depth in the realm of privacy.
At some point I'll probably try switching to reading RSS feeds from within Emacs, which should be even more flexible and powerful.
What do you use for RSS? After Google Reader shut down, I assumed very few people continued using RSS.
[1] - https://newsboat.org/
[2] - https://github.com/newsboat/newsboat