Hacker News Reader

Comment by tester756 | Hacker News Reader

Despite all the hate Google and infinite scroll gets, it seems to be good decision doesn't it?

You rarely go to even the second page, I'm almost always improving/changing the query instead of going to next pages, so even in those edge cases where I have to go to next pages, then auto loading next results when I'm at the bottom of the page seems viable solution

Infinite scroll is annoying because e.g you cannot link to thing

but that's not the case with Google - you do not link Google results to people, you send them the result itself - desired website.

View on HN

My biggest problem with infinite scrolling in general is it clears scrolling state by browser back or page refresh.

Sometimes there's no good result in the first page, or I want to keep searching for more investigation.

When I find a seemingly interesting result after several scrolls and visit the page but find out it's not what I want, I want to go back to the search result and continue. But after browser back, the result is cleared and I have to scroll several times again. Even if I open the result in the different tabs and keep the search result, it may disappear after browser refresh or restart.

> it clears scrolling state by browser back

Twitter handles this masterfully (or used to)

> or page refresh.

Technically you can use history.replaceState to store, ahem, state on the "current page", so nothing stops you from saving a cursor there to use when loading the content.

The issue with infinite scroll is that very few people know how or care to "make it work correctly"

I wish browsers did this as a default option. The back button should never have a noticeable delay.

They do if you use non-ajax navigation: the scroll position is preserved and so is the document as last seen: https://web.dev/bfcache/

But once you call history.pushState, the browser is relinquishing control to you. At that point, going “back” only changes the URL, as far as the browser is concerned, and that part is instantaneous.

I’ve taken to “open link in new tab” on any search result or listing page for years. It doesn’t solve all the issues, but makes things more tolerable.

> Infinite scroll is annoying because e.g you cannot link to thing

There are other reasons. It's annoying to me because on mobile there's a link I use often that's only in the footer of the page.

And yep, the link is still there in the infinite scroll version of the results page - if you swipe violently downward several times then after 3-4 refreshes it gives up, and you can click footer links.

(The link in question is "see results in English" - which I use often due to my region and because I use incognito tabs for random google searches.)

Have you tried adding lang:en to your query?

> Infinite scroll is annoying because e.g you cannot link to thing

Infinite scroll is also annoying because it often breaks when navigating back to the page.

Since this is a very popular use case in the context of a search page, I’m not a fan of this change.

It seems trivial for Google to fix if this is truly broken. They could just update the page params while scrolling.

In theory yes.

In practice when I go to youtube’s search page (safari mobile), scroll down past infinite scrolling, click a random video and go back, I don’t see the video that I clicked on my screen.

I haven’t seen the google’s search page with infinite scrolling. But even if it works now, I’d be worried that it would break in future, since so many popular websites don’t care about that.

In practice, it works if it’s implemented like on the desktop version of YouTube.

Maybe there’s a reason why popular websites don’t care?

Why “worry” about (possibly) expected features “breaking”? Seems extreme. Bugs happen. Sometimes they get fixed. Sometimes they don’t.

> You rarely go to even the second page

You might not, I do, I do a lot, I even go further. I personally don't like the change at all

But this saves you a click so what do you dislike?

With distinct pages it's easy to bookmark them and/or remember "oh, there was something interesting on page three." With infinite scroll, that interesting thing becomes lost in a maze of twisty passages and more difficult to return to because you are never on an individual static page, but rather a single ever-changing page.

If you're going through a long list of unchanging results, like issues of a magazine sorted by date, and you want to save your place, then sure.

But Google results are ephemeral and change constantly. I've never heard of anyone wanting to bookmark the second page of search results.

I agree infinite scroll doesn't make sense on all sites, but for internet search results it's perfect.

If you find an interesting search result on page 3 you seriously bookmark page 3 rather than bookmarking that link?

I'm curious - in what cases do you go even further?

In theory there might be more than ten interesting results on a topic, and with some search engines, say the one at

https://pubmed.ncbi.nlm.nih.gov/

there really are. For Google the results past the top ten have been unusable for a very long time, frequently just a repetition of the same garbage.

Circa 2004 I was interested in the question of why Google was so much better than other search engines, which got me into reading TREC proceedings which made it clear that Google probably wouldn't do very well on TREC because the original TREC scores assumes you care about the relevance of the 1000th result and Google was all about "I'm feeling lucky".

A month or so ago I was talking to someone developing a search engine who was concerned that most tracks in TREC now are concerned about precision @ 1, precision @ 5 and things like that and was wondering how to tune up a search engine if the 1000th result matters.

I guess things go full circle.

If your first language is English, and you end up using Google in a non-Anglophonic country, the option to change language (without going through sign-in hoops) is a drop-down at the bottom. The infinite scrolling, where results get lazy-loaded is really irksome in that case.

I used to often do a minimum of 5 pages when I was doing any sort of serious research to start checking the more esoteric and less SEO optimized results. The infinite scroll has definitely been annoying for me

In cases where the result wasn't on the previous page.

I can't decide what's more annoying though, the infinite scroll or the search result layout change.

What do you find preferable about clicking the next page link vs having it appear without a click?

Yup this is definitely an improvement for me.

Having to page through ephemeral search results is just added friction. I'm already scrolling anyways so this is much better UX.

That right there is the distraction though, more "efficient" for a corporate but worse for everybody else. Like you get a catch-22 but it's got its own catch-22. Issue ain't removing pages like you'd share page links, issue is making search "so efficient" clicking on page 5 or 10 direct is worthless then removing pages because google went and done made it worthless first. It's a buncha cattywampus and the real sad thing is Google only got better for itself you ask me, so it ain't really about sharing links to pages it's about pages having been made worthless

It’s not.

It makes sense when your incentive is to sell more ad impressions.

And the next logical step is to further reduce result quality so you have to scroll for more ad impressions.

It shows they are now treating the landing page as an ad revenue generator with no bounds, it’s no longer primarily a signals play.

Reply on news.ycombinator.com