Hacker News Reader

Comment by tokenadult | Hacker News Reader

Daniel Gackle ('gruseom) is an absurdly great choice as moderator.

Hear. Hear.

His submission list

https://news.ycombinator.com/submitted?id=gruseom

is a fine example of articles that "gratify intellectual curiosity." This leadership adjustment bodes well for the future of HN as a community for thoughtful discussion.

View on HN

>> "His submission list https://news.ycombinator.com/submitted?id=gruseom is a fine example of articles that "gratify intellectual curiosity.""

While I agree with you it's interesting that only one submission in the last 84 days has had significant up votes. What does that say about the people on HN or the kind of stories they are looking to find here?

I think it speaks mostly about the state of the "new" section. Submitting a worthwhile article is a very fickle undertaking, skewed by timezones and the fact that few people visit the section, or remember to upvote after reading a long article.

I could imagine a variant of the endorsement system for submissions, where you can endorse users and have their submitted stories automatically get your vote. But that seems rather open to gaming.

I thought someone collected data and found that timezones don't matter too much (times where fewer people are reading also have fewer other submissions to compete with.)

It doesn't seem like many people visit the news feed at all. Personally I imagine the most important features are picking a popular article that multiple people will try to submit (each new submission upvotes the original article instead) and maybe people who have friends with HN accounts. The title probably also makes a large difference.

Some people have suggested improvements to HN's sorting algorithm which would favor more "exploration" than relying on people to go through the new section.[1]

[1] http://www.bayesianwitch.com/blog/2013/why_hn_shouldnt_use_r...

It doesn't seem like many people visit the news feed at all.

Yes, and thus the rare souls that venture there have a ridiculously disproportionate impact on what gets the front page. Almost makes you wonder if it would make sense to keep access to the 'new' page as a special privilege for experienced and trusted users. Or at least hide it behind some tiny text link so those newbies don't ruin the party!

I upvoted a story on the new feed about the Inmarsat's approach to estimating the missing Malaysian plane's location that I found highly informative. Hours later that day when I noted the usual Telegraph mc-article rising up, not a single upvote on the better article(IMO) had been recorded. Perhaps it was a time delay as there is 1 upvote now. Shame, it seemed like the perfect article to elicit meaningful discussion on HN. Maybe Part 2 or Part 3 will someday...if they're proven correct, which is up for debate still.

HN post: https://news.ycombinator.com/item?id=7430634

DirectLink http://www.airtrafficmanagement.net/2014/03/mh370-satcoms-10...

I find that a comment as well as an upvote helps more. I'm more likely to check out a new article that has a comment, as there may be some discussion to be had, and I find the discussion the best part of the site.

Good point. I bookmarked that one due to length and had to revisit it later when time allowed. I checked back to see how it's HN status fared and read some additional insights, the discussions often add new perspectives or are more substantive than the articles themselves.

Checking the HN comments to ensure I don't inadvertently benefit a click-baiter is very useful, too.

One issue I think is that if there are two articles, one brilliant and one so so but with a catchy title they may both get one or two up votes so there is no way to tell at a glance which is the outstanding one. A possible way around that would be on the new stuff to have some way to vote as to the quality - say a score out of 10 or buttons reading 'so so' 'really interesting' and the like. They could make a tag pop up saying 'interesting' or similar but not actually effect the ranking on the page which would work on vote count as it does at the moment. I just mentioned that because I'm working on building a similar voting system myself as a kind of holiday project.

Maybe articles could have links to "similar" articles, or if you post a link which is similar, you're asked if you want to comment in the existing thread instead?

It seems like the problem might be that there's such a short amount of time to discover good content that gets submitted. During peak times, a submission can be pushed off the first page of /newest within ~20 minutes. If that's a problem, then one solution might be to change the number of items displayed on the new page from 30 to 120.

I wouldn't be surprised if the first page of /newest gets 5x to 10x more views than the second page. People like scrolling; people click much less frequently. So upping the number of posts shown on the new page to 120 should give better content more time to be discovered.

Another idea is to tweak the new page so that the first page of /newest shows every new submission for exactly two hours. That is, page 1 would consist of submissions that are 1 minute to 2 hours old.

Also, maybe some kind of reminder to check the new page every so often would help, but that's probably a bit heavy-handed.

> I wouldn't be surprised if the first page of /newest gets 5x to 10x more views than the second page.

My guess is way way less (not that /newest gets any traffic as it is). I regularly visit /newest to contribute a bit of voting, and I don't think I've ever bothered with the second page.

There's no point in visiting the second page. Even if you upvote something there, it's the only upvote it will get.

One of the things I sort of hate is when ~10 articles on the same topic get posted at the same time -- honestly, WaPo and NYT and CNN might not be blogspam about a press release or major news event, but there's often little need to read all of them (or rather, enough overlap that the best article would be a synthesis).

HN stories being anchored by a single URL kind of makes it impossible to put all the topical articles in one big story.

You're just explaining variance. But since no more than a handful of his 100+ posts made it to the front page, it seems pretty clear that they are not what HN selects for.

Not necessarily. Variance is a powerful force. 100+ samples may sound like a lot, but it's often insufficient. My favorite book on this topic is http://www.statisticsdonewrong.com/index.html, in particular the sections on statistical power.

There's a great deal of randomness with this stuff, so one has to be careful about drawing conclusions. If you guys look back further you'll see I've had more than my share of hitting the front page with things that are interesting and off the beaten track.

That said, we do have some ideas for counteracting that randomness, and it would be sweet if they work. I'd also be pleased to hear any suggestions on this. HN's (semi-)openness to interesting historical and literary material is one of my favorite things about it, and as long as I'm doing this job that won't change.

My favorite, by the way, of all the things I posted recently that didn't get any attention, is https://news.ycombinator.com/item?id=6713259.

>My favorite, by the way, of all the things I posted recently that didn't get any attention, is https://news.ycombinator.com/item?id=6713259.

Off topic. So, what's the opinion on RD Laing currently in the scientific community? I remember reading "A divided self" and wondering the whole time, is the whole world mad? Or is the author mad? Or is it both?

Wow, those are great posts. Pity I never saw most of them on the front page...

+1 on this. Impressed by the submission list.

Reply on news.ycombinator.com