Most-favorited Hacker News posts • Hacker News Board

Re: Most-favorited Hacker News posts

#261

Post by dang » Thu, Sep 03, 2020, 5:21 PM UTC

losvedir wrote: post #252

dang wrote: post #235

Earlier quoted context omitted.

140 https://news.ycombinator.com/item?id=4247615 138 https://news.ycombinator.com/item?id=15603013 108 https://news.ycombinator.com/item?id=18442941 93 https://news.ycombinator.com/item?id=13436420 86 https://news.ycombinator.com/item?id=8902739 81 https://news.ycombinator.com/item?id=11042400 81 https://news.ycombinator.com/item?id=14948078 76 https://news.ycombinator.com/item?id=6199544 65 https://news.ycombinator.…

Ha, I see three of those are already gray-visited for me, and clicking them now, I see why! * https://news.ycombinator.com/item?id=18442941 was a great look at Oracle DB development * https://news.ycombinator.com/item?id=35079 was one of the most hilarious situations of all time on HN * https://news.ycombinator.com/item?id=9224 was a fun early Dropbox "No wireless, less space than a Nomad, lame" moment. I'm gonna hav…

Re 9224, I feel bad that BrandonM has been singled out that way because the comment has been universally misinterpreted. He was commenting on Dropbox's YC application (that's what "app" meant back then) and trying to be helpful.

https://news.ycombinator.com/item?id=23229275

Re: Most-favorited Hacker News posts

#262

Post by dang » Thu, Sep 03, 2020, 5:26 PM UTC

amelius wrote: post #246

I'd like to see collaborative filtering applied to this. https://en.wikipedia.org/wiki/Collaborative_filtering

You mean "people who favorited X also favorited Y"? That could be interesting. Not sure there's enough data.

Re: Most-favorited Hacker News posts

#263

Post by dang » Thu, Sep 03, 2020, 5:28 PM UTC

ekianjo wrote: post #36

Not very useful since almost no-one uses the favorite function on HN.

68k users have used the favorite feature, on 1.7M posts. It's not widely used by HN standards, but that's certainly not "almost no-one".

You can also gauge it by how many people rushed to this thread to post "please don't remove the feature" when someone mentioned we had thought about removing it.

Re: Most-favorited Hacker News posts

#264

Post by fluffything » Thu, Sep 03, 2020, 5:44 PM UTC

ffpip wrote: post #59

fluffything wrote: post #4

Earlier quoted context omitted.

I use the mobile app and its horrible. I can't often save comments, i can't see my comment threads, ... There are probably many HN features nobody knows about because the UI doesn't expose them.

Which on are you using? There are many. I use this. It is very good imo - https://play.google.com/store/apps/details?id=io.github.hidr...

I'm using one called Hacker News (YC) from Dharmesh Patel. I thought that was the official app.

Re: Most-favorited Hacker News posts

#265

Post by nitrogen » Thu, Sep 03, 2020, 6:23 PM UTC

toomuchtodo wrote: post #144

abnry wrote: post #112

Earlier quoted context omitted.

Hide is important to me too. I don't want to be worked up by the threads that I know will br charged that I'm tempted to click on.

Seconded. I hide threads I know I won’t be able to have a productive conversation in.

This may have been referring to the "hide" link on the HN homepage, rather than collapsing threads.

Re: Most-favorited Hacker News posts

#266

Post by TheRealPomax » Fri, Sep 04, 2020, 12:24 AM UTC

oefrha wrote: post #137

TheRealPomax wrote: post #132

Earlier quoted context omitted.

there are polls?

https://news.ycombinator.com/newpoll Hardly ever see it used in any capacity though. Re gp: removing buttons shown on every page is probably more about simplifying interface and reclaiming space than maintenance overhead.

there's no space to reclaim, honestly. If removing a 12pt line of text saves space, please buy a phone that supports more than just WAP.

Re: Most-favorited Hacker News posts

#267

Post by tlarkworthy » Fri, Sep 04, 2020, 11:07 AM UTC

comboy wrote: post #251

tlarkworthy wrote: post #186

Earlier quoted context omitted.

I did normalize by number of favorites too, so each person had 1 vote shared over all their favorites. The idea that a person who only had one favourite was a really strong signal for that single link, whereas people who have lots of favorites are using it differently to what I want to select for. At some point I had a similar selection of Ask HN dominance. One problem with my data is I only grabbed the first page of…

> The idea that a person who only had one favourite was a really strong signal for that single link Perhaps karma or account age should join the equation.

Yeah, more recent articles should be penalised to normalize for the increased quantity of accounts.

Re: Most-favorited Hacker News posts

#268

Post by ekianjo » Sat, Sep 05, 2020, 1:07 AM UTC

bad_user wrote: post #77

ekianjo wrote: post #38

Earlier quoted context omitted.

HN Comments is where the money is, not the stories themselves.

Comments have links that can be bookmarked too, under that timestamp, here's yours: https://news.ycombinator.com/item?id=24352253

I usually favorite a whole story so that I have access to all comments, not just one.

Re: Most-favorited Hacker News posts

#269

Post by ekianjo » Sat, Sep 05, 2020, 1:14 AM UTC

dang wrote: post #263

ekianjo wrote: post #36

Not very useful since almost no-one uses the favorite function on HN.

68k users have used the favorite feature, on 1.7M posts. It's not widely used by HN standards, but that's certainly not "almost no-one". You can also gauge it by how many people rushed to this thread to post "please don't remove the feature" when someone mentioned we had thought about removing it.

Follow up questions when you want to measure adoption:

- 68k users out of how many users on HN? This is the metric that really matters when you want to consider how wide the adoption is.

- Number of users may not be very easy to evaluate as there's throwaway accounts, inactive accounts and the like, so ideally you would need to account for the people who log on at least once every month or something.

- Among those, you would need to filter those who actually use the feature regularly vs the ones who just used the feature once and never used it again. Look like the average is at 25 per user based on the data you provided, but I am guessing there are users with 100s of favorites so that means there are people at the end of the tail as well, with just a few, which should not be counted as "active favorite users".

By the way @dang, is there any hidden function on HN to see some HN related stats?

Re: Most-favorited Hacker News posts

#270

Post by ekianjo » Sat, Sep 05, 2020, 1:17 AM UTC

dang wrote: post #197

pc86 wrote: post #40

Earlier quoted context omitted.

As evidenced by the top favorite among the 10,000 most active users having less than 100, when it's over four years old and has nearly 1700 upvotes.

"Most active" in this case means "posting the most comments". That's an interesting subset to look at but there's no reason it would be correlated with heavy use of the favorites feature.

I agree, the majority of people who visit content aggregators tend to be readers, not commenters, and therefore there is no necessary correlation between commenting and bookmarking.