Hacker News Reader

Comment by hollerith | Hacker News Reader

I don't like how the home page at http://news.ycombinator.com/ sorts by a combination of score and time-since-posting because that means that I end up scanning each entry (story title) over and over.

In contrast, if they were sorted strictly by time-since-posting, I would probably quickly be able to tell once my scan of the list reached the part of the list I had already seen.

Someone else (many years ago) had the same thought as what I wrote above, motivating him to create http://hckrnews.com/, which is mainly what I use to find stories to read. (I also use https://hn.algolia.com/?query=%s&sort=byDate to find stories to read.)

View on HN

> if they were sorted strictly by time-since-posting

Isn't that exactly what you get on the “New” page?

https://news.ycombinator.com/newest

That page shows all stories whereas hckrnews.com shows only stories that are or were on HN's front page.

Ah. But... If you don't want to trust the front page's ordering of items, then why trust its choice of which to include? Seems a bit weird to me. Or, well, at least arbitrary.

It has nothing to do with trust.

OK, so let's use another word: If you're satisfied with letting HN's front-page algorithm select which posts you even get to see, then why not with the much smaller matter of ordering? (But of course it does: You being OK with an algorithm doing something for you is just another way of saying you trust the algo to do it for you.)

Anyway, trivial problem: The post IDs in the URLs are monotonally increasing over time. Just sort by ID and the lowest is the oldest, the highest the newest.

>why not with the much smaller matter of ordering?

I explained already, but am happy to explain again, adding a little background information. I need to keep to an absolute minimum the amount of time I spend looking for generally-interesting things to read on the internet.

Why? Because I have learned the hard way that I need to minimize the amount of time I spend doing things that are pleasurable and effortless and absolutely safe (not in the least dangerous). Consequently, if the project I am currently engaged in is to learn more about oxidized cholesterol, for example, I give myself permission to use hn.algolia.com to search HN for that phrase and read all I want. In contrast, I strictly limit the amount of time I use HN just to see if there are any interesting new stories when I don't have a definite goal for my reading.

hckrnews.com differs from the standard HN home page in that it gives me an easy way to tell when there are no headlines I haven't read already. Specifically, after reading 1 or 2 or 4 headlines I've already read, I know I've seen all the headlines below those 1 or 2 or 4 headlines, too, because the order never changes. That means that most headlines I only read once.

The order hckrnews.com or news.ycombinator.com chooses to sort the headlines is immaterial to me. What matters is that the order does not constantly change.

And that is my answer to "How do you browse Hacker News?" (the headline of the OP).

Ah, I see.

Then I'd recommend you use HN's “new” page instead. It gives you that exact advantage you seek, of knowing that when you recognise a headline (perhaps by it showing up as “followed” in your browser), then you've seen the ones after (i.e. chronologically before) it, and it gives you the chance to see the ones that didn't make the front page.

Now sure, the HN algo really isn't all that bad; if you do this, you'll get to scan past a lot of dross. But every once in a while, there's a pearl in amongst the muck, some quirky bit of stuff that speaks to you, even though it didn't to almost all of the readership whose input the algorithm is based on.

That's what I do (when I'm at my most compulsive).

Or you could press hide once you've seen a story.

I sometimes use hckrnews.com, but not for that reason - it's good for catching up on what's been on the front page if I haven't been here lately.

I hadn't considered that.

Reply on news.ycombinator.com