Live data from Hacker News

Infinite Scrolling: When to Use It, When to Avoid It

nngroup.com

41–50 of 175 posts

Re: Infinite Scrolling: When to Use It, When to Avoid It

#41
post #2

Infinite scrolling is universally bad for one, very simple reason: It breaks the scroll bar. One of the purposes of a scroll bar is to indicate how long (and/or wide) some piece of content is relative to your window. Another purpose is to indicate your current X/Y position within some piece of content. Both of those purposes are completely broken with the use of infinite scrolling, because the length (and/or width) o…

On one hand, yes, I agree. On the other hand, it seems to me that this is partly the fault of the scrollbar as an idea. How do you show the current position in a document for which, for one reason or another, you cannot instantly compute the visible height? I still remember an Android PDF viewer which would run through every page of a 2000-page document to get the page dimensions before displaying anything. (From wha…

I remember Android apps some years back that showed an approximate scrollbar, when it knew how many items there were (paragraphs, in the case I’m vaguely recalling—and although all the content was available, it might not have wanted to render it all at once, as it could have tens of thousands of words) but not how tall each was. I think it drew a scrollbar with the assumption that off-screen items would be a certain height, so if you went from a screenful of long paragraphs to a screenful of short paragraphs the scrollbar could be amusingly wonky, considering our usual expectation of spatial scrollbars. In some contexts this itemwise rather than spatial scrollbar can actually be more useful, but whether or no, it’s not a concept the web supports directly, and there will be certain niggles in the result of implementing it.

Re: Infinite Scrolling: When to Use It, When to Avoid It

#42
post #11
post #2

Infinite scrolling is universally bad for one, very simple reason: It breaks the scroll bar. One of the purposes of a scroll bar is to indicate how long (and/or wide) some piece of content is relative to your window. Another purpose is to indicate your current X/Y position within some piece of content. Both of those purposes are completely broken with the use of infinite scrolling, because the length (and/or width) o…

It also breaks the "virtual scrollbar". if you reload the page you have to start all over

This is my primary issue with infinite scrolling. However, I wonder if it’s mostly a problem with today’s web browsers.

Heck—maybe it’s not even browsers, but websites. I’m pretty sure a website could change the URL as you scroll, so if the page is reloaded (or bookmarked, etc) you’re brought back to where you left off. I usually dislike it when websites mess with URLs, but this seems like an excellent use case!

Re: Infinite Scrolling: When to Use It, When to Avoid It

#43
post #38

Earlier quoted context omitted.

I use the scrollbar on mobile. It's a way to quickly navigate to a section of a particularly large document that would take tons of swiping. Not a frequent use case but an absolute necessity on occasion.

Interesting, none of my mobile browsers even have a scroll bar.

An autohiding scroll bar is still a scroll bar.

Re: Infinite Scrolling: When to Use It, When to Avoid It

#44
post #11

Earlier quoted context omitted.

It also breaks the "virtual scrollbar". if you reload the page you have to start all over

This is my primary issue with infinite scrolling. However, I wonder if it’s mostly a problem with today’s web browsers. Heck—maybe it’s not even browsers, but websites. I’m pretty sure a website could change the URL as you scroll, so if the page is reloaded (or bookmarked, etc) you’re brought back to where you left off. I usually dislike it when websites mess with URLs, but this seems like an excellent use case!

even if the website did re-start where you left off (difficult), what if you scroll backwards? When you send a link, should it point to the page that you scrolled to, or to the beginning of the list? How many items before/after should load? These arise solely because people chose to reinvent a browser -in-the-browser. With pagination we dont have these issues, the link points to a specific results page, uniquely.

Re: Infinite Scrolling: When to Use It, When to Avoid It

#48
post #17

Earlier quoted context omitted.

> How do you show the current position in a document for which, for one reason or another, you cannot instantly compute the visible height? Perhaps by not "instantly" displaying the document- that is, displaying it before you are able to display its scrollbar.

Could someone remind me what the name is for the eternal debate between the two valid camps: “HTML should never ever render if there’s an error” vs. “Renderer should try its best to render part of a document if there’s errors” ? I feel that this is akin to that. I strongly believe both perspectives are valid for often overlapping contexts.

Draconian error handling vs Postel’s law? Both behaviours are still in browsers for XHTML (and XML+XSLT) resp. HTML (though the removal of in XHTML5 makes me sad).

Re: Infinite Scrolling: When to Use It, When to Avoid It

#49
My confidence in the guidance from this article was shaken when it talked about “Increased page load” and “Poor SEO performance” issues.

Their points are wrong and border on naivety.

Page load issues can and should be nonexistent fire to loading via xmlhttprequest/fetch ahead of time, and SEO issues are minor and typically negated by sitemap XML.

There are plenty of other best-practices to address lingering concerns.

Re: Infinite Scrolling: When to Use It, When to Avoid It

#50

Infinite scrolling makes the contents of the page unlinkable, breaks searchability and caching. It always makes the experience worse for unclear benefits. What's the problem it solves? Not resending the header and the footer on changing pages? Really?

No. Only if you do a bad job of implementing it. All those issues are avoidable.
Post reply on HN