Hacker News Reader

Comment by stubish | Hacker News Reader

We have this on one of our web apps, and it is a problem. To me the problem is why do we allow jumping to page 107 of 736 in the first place? How does this help someone? The two use cases are 1) user is manually performing a binary search on the results, 2) a bot needs to index the site. Both of these use cases are better solved in other ways, with either search forms that do not suck or index pages generated specifically for crawlers. So the answer to 'how would you implement a deep pagination like "go to page 107"' is to not implement that, and instead improve your design.

View on HN

>user is manually performing a binary search on the results

Imagine a large forum topic. The user is just seeking through time. A deep pagination here is when the user knows what they're searching for is more than 1 page away.

The user would have to know a snippet of text or an explicit date range for your search box to help, so what about all the times they don't or if their filters are insufficient? This is why forums have both deep pagination and a search box.

It would be cruel to make the user paginate one page at a time just like it'd be unbearable if Youtube didn't let you jump 10 minutes at a time.

Obviously not every pagination needs deep pagination, but that's between you and your users, and it's what came to mind when I read that article.

And that’s why Google’s pagination on search results is broken, and why so much usability was lost. Sometimes I want the lowest ranked results for a query, or the middle-of-the-field results. Let me have them.

>Sometimes I want the lowest ranked results for a query, or the middle-of-the-field results.

What's the use case for that?

Google results are gamed so much, that often personal blogs that don’t do SEO, but offer interesting content, appear beyond page 5 of Google search results, but are still relevant. Similar issues appear in many situations where search is used on third party results, but even on first-party results the first page is often gamed.

I’ve been thinking about it, and ideally one would make a search engine that only indexes pages that have no analytics, tracking, ads, paywalls, etc. Then I’d find those same pages I’m searching for.

To give an example: I found http://blog.deconinck.info/post/2016/12/19/A-Dirt-Cheap-F-Aw... on page 3 of Google for "raspberry pi led", just below https://tech.scargill.net/home-control-2016/ (both of which are interesting IMO)

It's a first result if you search for "raspberry pi led table"...

Yes, but I wasn't looking about tables. Just general things to do with led strips and raspberry pis.

Google is great for finding things you already know, but I'm using it for discovering things I didn't even know I was interested in.

I can't exactly just put the entire dictionary into google to try and find what I might like. I enter a category I might like, and look through the results matching that category.

Reply on news.ycombinator.com