Leapstories

Has anything on page 2 of YC Hacker News ever bothered you?

The longer you are on page 1, the more page 2 degrades. By the time you click `more`, some of the stories that were previously on page 2, or lower have snuck through to page 1. You never see them. The newer and more popular the story, the more likely it is to "leap".

This extension captures any rising stories that leapt into the existing page while you were browsing and seamlessly prepends them to the top of the next page as it loads. That's it.

### How it works

When you load a hackernews page, an array of the story ids are snapshot locally. When you click more, the current page is fetched behind the scenes and diffed with your snapshot. Any new stories are placed at the top of page 2 (or whatever the next page number is). Any stories on this next page that had already been seen on the previous page are optionally indicated as such.

### Why?

This isn't meant to be a completionist style extension to make sure you see all of the stories all of the time. This is meant to fill a gap in pagination strategy for this particular use case. A consistent, point-in-time, query or live polling would also address the issue. They're more complicated to implement and would require significantly more request load on hacker news when implemented client-side as a web extension. Neither are truly necessary.

This extension provides a nice balance between a static, sorted list while capturing just enough of the dynamic nature of the ranking in order to ensure that stories are seen at least once.

### Scope and Accuracy

Currently targets only the top stories listing (/ and /news).

Only page 1 to page 2 is fully accurate. On deeper pages, only the previous page is checked, rather than all previous pages. Stories that were injected from the page before that are not tracked. This means that if a story jumped multiple pages at once, or after being injected in a page, moved down another page, it will not be captured. This is intentional to reduce complexity and keep load on hacker news as light as possible.

The extension intentionally introduces minimal to no additional styling and doesn't try to limit or be smart about when or how many stories are pulled into the next page. If you click "more" after a day or two, you will see the near entirety of the current previous page. Read through if you want, or use this as a sign that you should just go back to the previous page.

### Permissions and Security

This uses as few permissions as possible. It only GETS from hackernews and, optionally for update checking, the github repo for this extension. No other HTTP requests. No tracking. No analytics. All data is local.

### Updates

It checks github, at most daily, for new releases, but does not auto install (this is optional and can be disabled). This is so you can confidently turn off auto updating or even install locally and still opt in to updates. It ONLY checks for updates if and when the extension actually gets used on hacker news. Thus, it doesn't do unnecessary work in the background when you're not actively on hackernews. When there is an available update, it's as unobtrusive as possible. You won't notice unless you have the icon shown or deliberately open the settings to check.