page scrolling when frame-like structures are present
Categories
(Core :: DOM: UI Events & Focus Handling, enhancement, P3)
Tracking
()
People
(Reporter: Nick_Levinson, Unassigned)
Details
PageUp and PageDown vertically scroll and with a good overlap, but with one exception. When there's something that behaves like a frame at the top (or probably at the bottom but the left or right don't matter), the scrolling goes too far. It skips lines. When skimming an article, a reader might not realize that they never saw some of the lines of content.
Examples:
Two contrary examples, with apparent top frames and nonetheless proper scrolling:
http://web.archive.org/web20150204210120/http://www.latimes.com/local/
https://www.encyclopedia.com/history/modern-europe/british-and-irish-history/republicanism
I'm lately finding them on Google search engine results pages, the apparent frame becoming visible as such only when I scroll down, but not when the top of the results page is visible.
(All URLs are as accessed Jan. 3-4, 2019, except for Quora, accessed Jan. 13, 2019, and Google on various days.)
I think I see the problem occur more than I don't, when apparent frames are present. However, that may be a function of my choice of websites to visit.
Frames per se are deprecated in HTML, I think because of a security problem. But they're still common and we shouldn't ignore them because of deprecation.
It's possible that what look and behave like frames are not frames but other things. Different sites may be using different technologies to achieve the general effect. Perhaps JavaScript is in use; I have no idea and I'm just guessing.
What is wanted is for the browser to detect the height of the frame-like structures at top and bottom and then scroll a pageful based on a net height that is the viewport height minus the two frame-like heights, regardless of the technology behind the frame-like structures.
|
||
Updated•5 years ago
|
|
||
Comment 1•5 years ago
|
||
Hi Mike, I'd be interested in learning your thoughts on priority with perspective of web compatibility.
|
||
Comment 2•5 years ago
|
||
Hi Nick, could you help me understand the issue a little bit better?
For the first URL, https://www.independent.co.uk/news/uk/home-news/meet-britains-republicans-is-the-uk-ready-to-think-about-change-in-the-year-the-queen-turns-90-a6826256.html
I load the site, hit page down but don't experience any skipped lines. The same is true for the Quora page, so it's possible I misunderstand the steps to reproduce. Thanks.
|
Reporter | |
Comment 3•5 years ago
|
||
A recent upgrade of Firefox to version 65.0 (64-bit) (fedora - 1.0) (I keep my laptop evergreen and rolling back is not feasible) seems to have cured the problem for the Independent and Quora pages (as accessed 2-5-19) but not for two which I'm adding below. Perhaps FF is newly coping properly with some technologies but still not all of them; perhaps those websites happened to have been rewritten recently.
Also, with the independent.co.uk example, when I opened ("+") the bottom thing promoting subscribing, a close button appeared; when I clicked that, the problem no longer existed. That's an ambivalent usability problem not worth including here.
More cases:
Standard problem: http://nymag.com/intelligencer/2019/02/northam-cites-time-he-wore-blackface-to-deny-racist-photo.html (as accessed 2-3-19 & 2-6-19).
In this one, the top frame-like structure conceals text only on the first page-down, not on subsequent page-downs and not when doing page-up back up: https://weather.com/science/news/2018-12-21-government-shutdown-what-it-means-weather-science-agencies (as accessed 2-5-19 & 2-6-19). The skipped line says "Congress failed to avert a partial government shutdown late Friday." If you page-down and want to try it again, go all the way back to the top and then page-down again.
Some EU/GDPR cookie notices have this effect until the user responds. Also, some of them don't occupy the full viewport width.
When the mouse cursor happens to be over a link, the URL displays as a tooltip at the bottom. Since that can happen even though a user does not intentionally line up the link text with the cursor when scrolling through a page, the net viewport height should deduct for the height of a one-line URL tooltip.
(The STR is as you and I said; you got it right.)
|
|
||
Updated•5 years ago
|
|
|
||
Updated•5 years ago
|
|
|
Reporter | |
Comment 4•5 years ago
|
||
Possibly, and this is just speculative, a browser's viewport width may affect what can't be seen due to a page-down. I'm using a laptop with display resolution 1440x900. I've not tried a cell pohone (my cell's browser is lousy). Perhaps, at some widths, whether content is even missed at all differs.
More examples (I may keep adding examples in case website tech differs):
On this one, instead of a frame-like object at the top of the viewport, it's near the top of the viewport: https://www.bloomberg.com/news/articles/2019-02-08/bezos-allegations-could-upend-american-media-s-deal-with-feds (as accessed 2-8-19). Another frame-like object is at the bottom but if a user is a Bloomberg subscriber they may not get that object. The hidden lines near the top on the second page-down: "supposed to refrain from all illegal activity for a three-year"/"period. The agreement says that if New York-based AMI"
https://www.deepintent.com/dsp (as accessed 2-8-19). On the first page-down, the line not seen without scrolling back up is "DeepIntent uses state-of-the-art clinical data and artificial intelligence to activate clinically-".
Usually ok but sometimes not: https://www.yahoo.com/ (home page, with news summaries) (as accessed 2-9-19):
--- Intermittency; usually nothing is missing but sometimes it is and I haven't figured out how to predict it. It may be on the first trip down after the first two page-downs. This might be a website-specific feature to help a visitor accellerate your way down, but, if so, it's interesting if a website can intercept a page-down and adjust the travel distance.
--- The top-of-viewport object has two heights, both generally compatible with not missing content during page-down or page-up. The shorter height has the logo, a search box, and a Sign In link. The taller height has all that plus links for News, Finance, etc. I'm not sure when you get which height, but it might matter whether you're paging up or down and it might matter whether you've just opened the page or you've been scrolling lately.
Contra, basically okay:
No content missing because of top-of-page object: https://www.thedailybeast.com/bezos-investigators-question-the-brother-of-his-mistress-lauren-sanchez-in-national-enquirer-leak-probe (as accessed 2-8-19).
https://www.thedailybeast.com/alexandria-ocasio-cortez-to-stephen-colbert-i-give-zero-fucks (as accessed 2-10-19) has a top-of window object and a closeable bottom partial-width object that hides text during scrolling. When that bottom object is closed, scrolling does not miss text.
This one scrolls by proper amounts whether the bottom object is present or closed and absent, and it may be adjusting the amount of scrolling by whether the bottom object is present: https://www.cfr.org/backgrounder/europes-migration-crisis (as accessed 2-17-19).
|
Assignee | |
Updated•5 years ago
|
|
||
Comment 5•5 years ago
|
||
See bug 1547409. Moving webcompat whiteboard tags to project flags.
|
|
||
Updated•5 years ago
|
Only the first time the PageDown key is pressed.
From the second press of the PageDown key, it's fine.
https://www.cio.com/article/230935/hiring-the-most-in-demand-tech-jobs-for-2021.html
https://zdnet.co.kr/view/?no=20220113102340
https://www.ciokorea.com/news/221804
https://www.itworld.co.kr/howto/221919
Most similar types of Koean sites have this problem.
Chromium Browser is OK.
|
||
Updated•2 years ago
|
Description
•